diff --git a/result/per_paper/2105.06643/accuracy_efficiency.csv b/result/per_paper/2105.06643/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2105.06643/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2105.06643/accuracy_efficiency_traced.csv b/result/per_paper/2105.06643/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2105.06643/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2105.06643/components_architecture.csv b/result/per_paper/2105.06643/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d4c9280410958cce26b2e37b4256b68c6e1baa04
--- /dev/null
+++ b/result/per_paper/2105.06643/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Monash Time Series Forecasting Archive,"A comprehensive time series forecasting archive containing 20 publicly available datasets with equal and variable lengths, including real-world and competition datasets across varied domains.",proposed_here,,"We introduce the first comprehensive time series forecasting archive containing datasets of related time series, available at https://forecastingdata.org/. This archive contains 20 publicly available time series datasets..."
+.tsf File Format,"A new time series data storage format based on the Weka ARFF file format, designed to store meta-information and series-specific details in a non-redundant, flexible manner.",proposed_here,"Paynter et al., 2008; Löning et al., 2019","We introduce a new format to store time series data, based on the Weka ARFF file format (Paynter et al., 2008) and overcoming some of the shortcomings we observe in the .ts format used in the sktime time series repository (Löning et al., 2019). We use a .tsf extension for this new format."
+tsfeatures,"A feature extraction method for time series data, used to analyze characteristics of datasets and identify similarities/differences among series.",reused_cited,"Hyndman et al., 2020","We analyse the characteristics of different series to identify the similarities and differences among them. For that, we conduct a feature analysis using tsfeatures (Hyndman et al., 2020)..."
+catch22 Features,"A set of 22 time series features designed for classification tasks, used here for feature analysis of the datasets.",reused_cited,"Lubba et al., 2019","We analyse the characteristics of different series to identify the similarities and differences among them. For that, we conduct a feature analysis using ... catch22 features (Lubba et al., 2019) extracted from all series of all datasets."
+Baseline Forecasting Evaluation Framework,"A framework for evaluating baseline forecasting models (traditional univariate and global models) across eight error metrics, with publicly available forecasts and results.",proposed_here,,We evaluate the performance of a set of baseline forecasting models including both traditional univariate forecasting models and global forecasting models over all datasets across eight error metrics. The forecasts and evaluation results of the baseline methods are publicly available...
diff --git a/result/per_paper/2105.06643/computational.csv b/result/per_paper/2105.06643/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2383984b38675e129e257e3e7f4b6024d26d3379
--- /dev/null
+++ b/result/per_paper/2105.06643/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,not_reported
+num_devices,,,not_reported,not_reported
+training_cost,,,not_reported,not_reported
+training_batch_size,,,not_reported,not_reported
+training_steps_or_epochs,,,not_reported,not_reported
+precision,,,not_reported,not_reported
+inference_latency,,,not_reported,not_reported
+inference_throughput,,,not_reported,not_reported
+peak_memory,,,not_reported,not_reported
+flops_or_macs,,,not_reported,not_reported
+num_inference_samples,,,not_reported,not_reported
+params,,,not_reported,not_reported
+context_lengths_evaluated,,,not_reported,not_reported
+horizon_lengths_evaluated,,,not_reported,not_reported
+inference_batch_size,,,not_reported,not_reported
diff --git a/result/per_paper/2302.11939/accuracy_efficiency.csv b/result/per_paper/2302.11939/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..63535c3615ad3df638d5ebd988332a318604073c
--- /dev/null
+++ b/result/per_paper/2302.11939/accuracy_efficiency.csv
@@ -0,0 +1,98 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Weather,GPT4TS / One-Fits-All (Large),MSE,0.221,
+accuracy,Weather,GPT4TS / One-Fits-All (Small),MSE,0.297,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.342,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.318,
+accuracy,ETTh1,GPT4TS / One-Fits-All (Large),MSE,0.613,
+accuracy,ETTh1,GPT4TS / One-Fits-All (Small),MSE,0.552,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.596,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.915,
+accuracy,ETTh2,GPT4TS / One-Fits-All (Large),MSE,0.413,
+accuracy,ETTh2,GPT4TS / One-Fits-All (Small),MSE,0.451,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.499,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.462,
+accuracy,ETTm1,GPT4TS / One-Fits-All (Large),MSE,0.774,
+accuracy,ETTm1,GPT4TS / One-Fits-All (Small),MSE,0.614,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.628,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.797,
+accuracy,ETTm2,GPT4TS / One-Fits-All (Large),MSE,0.352,
+accuracy,ETTm2,GPT4TS / One-Fits-All (Small),MSE,0.454,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.930,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.332,
+accuracy,ECL,GPT4TS / One-Fits-All (Large),MSE,0.261,
+accuracy,ECL,GPT4TS / One-Fits-All (Small),MSE,0.348,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.478,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.444,
+accuracy,Traffic,GPT4TS / One-Fits-All (Large),MSE,0.672,
+accuracy,Traffic,GPT4TS / One-Fits-All (Small),MSE,0.405,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.446,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,1.453,
+accuracy,Average,GPT4TS / One-Fits-All (Large),MSE,0.559,
+accuracy,Average,GPT4TS / One-Fits-All (Small),MSE,0.674,
+accuracy,Weather,GPT4TS / One-Fits-All (Large),MSE,0.227,
+accuracy,Weather,GPT4TS / One-Fits-All (Small),MSE,0.299,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.353,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.327,
+accuracy,ETTh1,GPT4TS / One-Fits-All (Large),MSE,0.681,
+accuracy,ETTh1,GPT4TS / One-Fits-All (Small),MSE,0.570,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.598,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.943,
+accuracy,ETTh2,GPT4TS / One-Fits-All (Large),MSE,0.428,
+accuracy,ETTh2,GPT4TS / One-Fits-All (Small),MSE,0.468,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.489,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.470,
+accuracy,ETTm1,GPT4TS / One-Fits-All (Large),MSE,0.726,
+accuracy,ETTm1,GPT4TS / One-Fits-All (Small),MSE,0.578,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.620,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.857,
+accuracy,ETTm2,GPT4TS / One-Fits-All (Large),MSE,0.232,
+accuracy,ETTm2,GPT4TS / One-Fits-All (Small),MSE,0.322,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.433,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.341,
+accuracy,ECL,GPT4TS / One-Fits-All (Large),MSE,0.297,
+accuracy,ECL,GPT4TS / One-Fits-All (Small),MSE,0.367,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.404,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.627,
+accuracy,Traffic,GPT4TS / One-Fits-All (Large),MSE,0.795,
+accuracy,Traffic,GPT4TS / One-Fits-All (Small),MSE,0.481,
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.502,
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,1.526,
+accuracy,ETTm1,GPT2(3),MSE,0.017,
+accuracy,Avg,GPT2(3),MSE,0.105,
+accuracy,ETTm2,GPT2(3),MSE,0.017,
+accuracy,Avg,GPT2(3),MSE,0.084,
+accuracy,ETTh1,GPT2(3),MSE,0.043,
+accuracy,Avg,GPT2(3),MSE,0.173,
+accuracy,ETTh2,GPT2(3),MSE,0.039,
+accuracy,Avg,GPT2(3),MSE,0.141,
+accuracy,ECL,GPT2(3),MSE,0.080,
+accuracy,Avg,GPT2(3),MSE,0.207,
+accuracy,Weather,GPT2(3),MSE,0.026,
+accuracy,Avg,GPT2(3),MSE,0.056,
+accuracy,GPT(6),GPT4TS / One-Fits-All (6),F1,6,
+accuracy,EthanolConcentration,GPT4TS / One-Fits-All (6),Imputation Error,34.2,
+accuracy,FaceDetection,GPT4TS / One-Fits-All (6),Imputation Error,69.2,
+accuracy,Handwriting,GPT4TS / One-Fits-All (6),Imputation Error,32.7,
+accuracy,Heartbeat,GPT4TS / One-Fits-All (6),Imputation Error,77.2,
+accuracy,JapaneseVowels,GPT4TS / One-Fits-All (6),Imputation Error,98.6,
+accuracy,PEMS-SF,GPT4TS / One-Fits-All (6),Imputation Error,87.9,
+accuracy,SelfRegulationSCP1,GPT4TS / One-Fits-All (6),Imputation Error,93.2,
+accuracy,SelfRegulationSCP2,GPT4TS / One-Fits-All (6),Imputation Error,59.4,
+accuracy,SpokenArabicDigits,GPT4TS / One-Fits-All (6),Imputation Error,99.2,
+accuracy,UWaveGestureLibrary,GPT4TS / One-Fits-All (6),Imputation Error,88.1,
+accuracy,Average,GPT4TS / One-Fits-All (6),Imputation Error,74.0,
+accuracy,Yearly Quarterly,GPT4TS / One-Fits-All (6),SMAPE,13.531,
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,2.996,
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.786,
+accuracy,SMAPE,GPT4TS / One-Fits-All (6),SMAPE,10.100,
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,1.182,
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.890,
+accuracy,Monthly Others,GPT4TS / One-Fits-All (6),SMAPE,12.894,
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,0.933,
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.878,
+accuracy,SMAPE,GPT4TS / One-Fits-All (6),SMAPE,4.891,
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,3.302,
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,1.035,
+accuracy,Average,GPT4TS / One-Fits-All (6),SMAPE,11.991,
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,1.585,
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.851,
diff --git a/result/per_paper/2302.11939/accuracy_efficiency_traced.csv b/result/per_paper/2302.11939/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b3b7a8a38de9d7cd5d1fd5fa222ae1d47bcbc2ca
--- /dev/null
+++ b/result/per_paper/2302.11939/accuracy_efficiency_traced.csv
@@ -0,0 +1,98 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Weather,GPT4TS / One-Fits-All (Large),MSE,0.221,,1,2,14
+accuracy,Weather,GPT4TS / One-Fits-All (Small),MSE,0.297,,1,2,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.342,,1,6,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.318,,1,6,15
+accuracy,ETTh1,GPT4TS / One-Fits-All (Large),MSE,0.613,,1,7,14
+accuracy,ETTh1,GPT4TS / One-Fits-All (Small),MSE,0.552,,1,7,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.596,,1,11,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.915,,1,11,15
+accuracy,ETTh2,GPT4TS / One-Fits-All (Large),MSE,0.413,,1,12,14
+accuracy,ETTh2,GPT4TS / One-Fits-All (Small),MSE,0.451,,1,12,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.499,,1,16,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.462,,1,16,15
+accuracy,ETTm1,GPT4TS / One-Fits-All (Large),MSE,0.774,,1,17,14
+accuracy,ETTm1,GPT4TS / One-Fits-All (Small),MSE,0.614,,1,17,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.628,,1,21,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.797,,1,21,15
+accuracy,ETTm2,GPT4TS / One-Fits-All (Large),MSE,0.352,,1,22,14
+accuracy,ETTm2,GPT4TS / One-Fits-All (Small),MSE,0.454,,1,22,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.930,,1,26,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.332,,1,26,15
+accuracy,ECL,GPT4TS / One-Fits-All (Large),MSE,0.261,,1,27,14
+accuracy,ECL,GPT4TS / One-Fits-All (Small),MSE,0.348,,1,27,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.478,,1,31,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.444,,1,31,15
+accuracy,Traffic,GPT4TS / One-Fits-All (Large),MSE,0.672,,1,32,14
+accuracy,Traffic,GPT4TS / One-Fits-All (Small),MSE,0.405,,1,32,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.446,,1,36,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,1.453,,1,36,15
+accuracy,Average,GPT4TS / One-Fits-All (Large),MSE,0.559,,1,37,14
+accuracy,Average,GPT4TS / One-Fits-All (Small),MSE,0.674,,1,37,15
+accuracy,Weather,GPT4TS / One-Fits-All (Large),MSE,0.227,,2,2,14
+accuracy,Weather,GPT4TS / One-Fits-All (Small),MSE,0.299,,2,2,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.353,,2,6,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.327,,2,6,15
+accuracy,ETTh1,GPT4TS / One-Fits-All (Large),MSE,0.681,,2,7,14
+accuracy,ETTh1,GPT4TS / One-Fits-All (Small),MSE,0.570,,2,7,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.598,,2,11,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.943,,2,11,15
+accuracy,ETTh2,GPT4TS / One-Fits-All (Large),MSE,0.428,,2,12,14
+accuracy,ETTh2,GPT4TS / One-Fits-All (Small),MSE,0.468,,2,12,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.489,,2,16,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.470,,2,16,15
+accuracy,ETTm1,GPT4TS / One-Fits-All (Large),MSE,0.726,,2,17,14
+accuracy,ETTm1,GPT4TS / One-Fits-All (Small),MSE,0.578,,2,17,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.620,,2,21,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.857,,2,21,15
+accuracy,ETTm2,GPT4TS / One-Fits-All (Large),MSE,0.232,,2,22,14
+accuracy,ETTm2,GPT4TS / One-Fits-All (Small),MSE,0.322,,2,22,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.433,,2,26,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.341,,2,26,15
+accuracy,ECL,GPT4TS / One-Fits-All (Large),MSE,0.297,,2,27,14
+accuracy,ECL,GPT4TS / One-Fits-All (Small),MSE,0.367,,2,27,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.404,,2,31,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,0.627,,2,31,15
+accuracy,Traffic,GPT4TS / One-Fits-All (Large),MSE,0.795,,2,32,14
+accuracy,Traffic,GPT4TS / One-Fits-All (Small),MSE,0.481,,2,32,15
+accuracy,Avg.,GPT4TS / One-Fits-All (Large),MSE,0.502,,2,36,14
+accuracy,Avg.,GPT4TS / One-Fits-All (Small),MSE,1.526,,2,36,15
+accuracy,ETTm1,GPT2(3),MSE,0.017,,3,2,2
+accuracy,Avg,GPT2(3),MSE,0.105,,3,6,2
+accuracy,ETTm2,GPT2(3),MSE,0.017,,3,7,2
+accuracy,Avg,GPT2(3),MSE,0.084,,3,11,2
+accuracy,ETTh1,GPT2(3),MSE,0.043,,3,12,2
+accuracy,Avg,GPT2(3),MSE,0.173,,3,16,2
+accuracy,ETTh2,GPT2(3),MSE,0.039,,3,17,2
+accuracy,Avg,GPT2(3),MSE,0.141,,3,21,2
+accuracy,ECL,GPT2(3),MSE,0.080,,3,22,2
+accuracy,Avg,GPT2(3),MSE,0.207,,3,26,2
+accuracy,Weather,GPT2(3),MSE,0.026,,3,27,2
+accuracy,Avg,GPT2(3),MSE,0.056,,3,31,2
+accuracy,GPT(6),GPT4TS / One-Fits-All (6),F1,6,,4,2,0
+accuracy,EthanolConcentration,GPT4TS / One-Fits-All (6),Imputation Error,34.2,,6,2,18
+accuracy,FaceDetection,GPT4TS / One-Fits-All (6),Imputation Error,69.2,,6,3,18
+accuracy,Handwriting,GPT4TS / One-Fits-All (6),Imputation Error,32.7,,6,4,18
+accuracy,Heartbeat,GPT4TS / One-Fits-All (6),Imputation Error,77.2,,6,5,18
+accuracy,JapaneseVowels,GPT4TS / One-Fits-All (6),Imputation Error,98.6,,6,6,18
+accuracy,PEMS-SF,GPT4TS / One-Fits-All (6),Imputation Error,87.9,,6,7,18
+accuracy,SelfRegulationSCP1,GPT4TS / One-Fits-All (6),Imputation Error,93.2,,6,8,18
+accuracy,SelfRegulationSCP2,GPT4TS / One-Fits-All (6),Imputation Error,59.4,,6,9,18
+accuracy,SpokenArabicDigits,GPT4TS / One-Fits-All (6),Imputation Error,99.2,,6,10,18
+accuracy,UWaveGestureLibrary,GPT4TS / One-Fits-All (6),Imputation Error,88.1,,6,11,18
+accuracy,Average,GPT4TS / One-Fits-All (6),Imputation Error,74.0,,6,12,18
+accuracy,Yearly Quarterly,GPT4TS / One-Fits-All (6),SMAPE,13.531,,7,1,2
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,2.996,,7,2,2
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.786,,7,3,2
+accuracy,SMAPE,GPT4TS / One-Fits-All (6),SMAPE,10.100,,7,4,2
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,1.182,,7,5,2
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.890,,7,6,2
+accuracy,Monthly Others,GPT4TS / One-Fits-All (6),SMAPE,12.894,,7,7,2
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,0.933,,7,8,2
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.878,,7,9,2
+accuracy,SMAPE,GPT4TS / One-Fits-All (6),SMAPE,4.891,,7,10,2
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,3.302,,7,11,2
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,1.035,,7,12,2
+accuracy,Average,GPT4TS / One-Fits-All (6),SMAPE,11.991,,7,13,2
+accuracy,MASE,GPT4TS / One-Fits-All (6),SMAPE,1.585,,7,14,2
+accuracy,OWA,GPT4TS / One-Fits-All (6),SMAPE,0.851,,7,15,2
diff --git a/result/per_paper/2302.11939/components_architecture.csv b/result/per_paper/2302.11939/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b64302cbd4cff60c56ca07482927aeae05c15ed0
--- /dev/null
+++ b/result/per_paper/2302.11939/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Frozen Pretrained Transformer (FPT),"A pre-trained transformer model (e.g., GPT2) with frozen self-attention and feedforward layers, adapted for time series analysis through fine-tuning.",proposed_here,"Zhou et al., 2023","This model, known as the Frozen Pretrained Transformer (FPT), is evaluated through fine-tuning on all major types of tasks involving time series."
+Self-Attention Mechanism,A module in the transformer architecture that captures long-range dependencies in time series data by computing attention weights across all positions.,reused_cited,"Vaswani et al., 2017","The self-attention module behaviors similarly to principle component analysis (PCA), an observation that helps explains how transformer bridges the domain gap."
+Feedforward Layers,Non-linear transformations applied to the output of self-attention mechanisms in each residual block of the transformer.,reused_cited,"Vaswani et al., 2017",We refrain from altering the self-attention and feedforward layers of the residual blocks in the pre-trained language or image model.
+Residual Blocks,Structural units in the transformer that combine self-attention and feedforward layers with skip connections to enable deeper network training.,reused_cited,"Vaswani et al., 2017",We refrain from altering the self-attention and feedforward layers of the residual blocks in the pre-trained language or image model.
+Fine-Tuning Mechanism,"A process of adapting the pre-trained model's parameters (excluding frozen layers) to specific time series tasks (e.g., forecasting, classification).",reused_cited,"Devlin et al., 2019",This model [...] is evaluated through fine-tuning on all major types of tasks involving time series.
+PCA Analogy for Self-Attention,A theoretical insight comparing the self-attention module's behavior to principal component analysis (PCA) for dimensionality reduction in time series.,proposed_here,"Zhou et al., 2023",We also found both theoretically and empirically that the self-attention module behaviors similarly to principle component analysis (PCA).
diff --git a/result/per_paper/2302.11939/computational.csv b/result/per_paper/2302.11939/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2302.11939/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2310.01728/accuracy_efficiency.csv b/result/per_paper/2310.01728/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..39cbc3525fd9b652adb8d1e0e1d159ff46d31fc1
--- /dev/null
+++ b/result/per_paper/2310.01728/accuracy_efficiency.csv
@@ -0,0 +1,90 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,Time-LLM,MSE,0.613,
+accuracy,Avg,Time-LLM,MSE,0.596,
+accuracy,ETTh2,Time-LLM,MSE,0.413,
+accuracy,Avg,Time-LLM,MSE,0.499,
+accuracy,ETTm1,Time-LLM,MSE,0.774,
+accuracy,Avg,Time-LLM,MSE,0.628,
+accuracy,ETTm2,Time-LLM,MSE,0.352,
+accuracy,Avg,Time-LLM,MSE,0.930,
+accuracy,Weather,Time-LLM,MSE,0.221,
+accuracy,Avg,Time-LLM,MSE,0.342,
+accuracy,Electricity,Time-LLM,MSE,0.261,
+accuracy,Avg,Time-LLM,MSE,0.478,
+accuracy,Traffic,Time-LLM,MSE,0.672,
+accuracy,Avg,Time-LLM,MSE,0.446,
+accuracy,ETTh1,Time-LLM,MSE,0.681,
+accuracy,Avg,Time-LLM,MSE,0.598,
+accuracy,ETTh2,Time-LLM,MSE,0.428,
+accuracy,Avg,Time-LLM,MSE,0.457,
+accuracy,ETTm1,Time-LLM,MSE,0.726,
+accuracy,Avg,Time-LLM,MSE,0.620,
+accuracy,ETTm2,Time-LLM,MSE,0.232,
+accuracy,Avg,Time-LLM,MSE,0.433,
+accuracy,Weather,Time-LLM,MSE,0.227,
+accuracy,Avg,Time-LLM,MSE,0.353,
+accuracy,Electricity,Time-LLM,MSE,0.297,
+accuracy,Avg,Time-LLM,MSE,0.404,
+accuracy,Traffic,Time-LLM,MSE,0.795,
+accuracy,Avg,Time-LLM,MSE,0.502,
+accuracy,$ETTh1 \rightarrow ETTh2$,Time-LLM,MSE,0.400,
+accuracy,Avg,Time-LLM,MSE,0.380,
+accuracy,$ETTh1 \rightarrow ETTm2$,Time-LLM,MSE,0.357,
+accuracy,Avg,Time-LLM,MSE,0.314,
+accuracy,$ETTh2 \rightarrow ETTh1$,Time-LLM,MSE,0.555,
+accuracy,Avg,Time-LLM,MSE,0.565,
+accuracy,$ETTh2 \rightarrow ETTm2$,Time-LLM,MSE,0.336,
+accuracy,Avg,Time-LLM,MSE,0.325,
+accuracy,$ETTm1 \rightarrow ETTh2$,Time-LLM,MSE,0.415,
+accuracy,Avg,Time-LLM,MSE,0.439,
+accuracy,$ETTm1 \rightarrow ETTm2$,Time-LLM,MSE,0.314,
+accuracy,Avg,Time-LLM,MSE,0.296,
+accuracy,$ETTm2 \rightarrow ETTh2$,Time-LLM,MSE,0.391,
+accuracy,Avg,Time-LLM,MSE,0.409,
+accuracy,$ETTm2 \rightarrow ETTm1$,Time-LLM,MSE,0.490,
+accuracy,Avg,Time-LLM,MSE,0.568,
+accuracy,ETTh1,Time-LLM,MSE,0.933,
+accuracy,Avg,Time-LLM,MSE,0.656,
+accuracy,ETTh2,Time-LLM,MSE,0.390,
+accuracy,Avg,Time-LLM,MSE,0.532,
+accuracy,ETTm1,Time-LLM,MSE,1.091,
+accuracy,Avg,Time-LLM,MSE,0.697,
+accuracy,ETTm2,Time-LLM,MSE,0.435,
+accuracy,Avg,Time-LLM,MSE,0.634,
+accuracy,Weather,Time-LLM,MSE,0.255,
+accuracy,Avg,Time-LLM,MSE,0.448,
+accuracy,Electricity,Time-LLM,MSE,0.520,
+accuracy,Avg,Time-LLM,MSE,0.510,
+accuracy,Traffic,Time-LLM,MSE,1.068,
+accuracy,Avg,Time-LLM,MSE,0.765,
+accuracy,ILI,Time-LLM,MSE,4.909,
+accuracy,Avg,Time-LLM,MSE,1.346,
+accuracy,Yearly,Time-LLM,SMAPE,13.419,
+accuracy,MASE,Time-LLM,SMAPE,3.565,
+accuracy,OWA,Time-LLM,SMAPE,0.911,
+accuracy,Quarterly,Time-LLM,SMAPE,10.110,
+accuracy,MASE,Time-LLM,SMAPE,1.253,
+accuracy,OWA,Time-LLM,SMAPE,0.938,
+accuracy,Monthly,Time-LLM,SMAPE,12.980,
+accuracy,MASE,Time-LLM,SMAPE,1.003,
+accuracy,OWA,Time-LLM,SMAPE,0.931,
+accuracy,Others,Time-LLM,SMAPE,4.795,
+accuracy,MASE,Time-LLM,SMAPE,4.116,
+accuracy,OWA,Time-LLM,SMAPE,1.259,
+accuracy,Average,Time-LLM,SMAPE,11.983,
+accuracy,MASE,Time-LLM,SMAPE,1.808,
+accuracy,OWA,Time-LLM,SMAPE,0.94,
+accuracy,ETTh1,Time-LLM,MSE,0.556,
+accuracy,ETTh1,Time-LLM,MSE,0.522,
+accuracy,ETTh2,Time-LLM,MSE,0.370,
+accuracy,ETTh2,Time-LLM,MSE,0.394,
+accuracy,ETTm1,Time-LLM,MSE,0.404,
+accuracy,ETTm1,Time-LLM,MSE,0.427,
+accuracy,ETTm2,Time-LLM,MSE,0.277,
+accuracy,ETTm2,Time-LLM,MSE,0.323,
+accuracy,Weather,Time-LLM,MSE,0.234,
+accuracy,Weather,Time-LLM,MSE,0.273,
+accuracy,ECL,Time-LLM,MSE,0.175,
+accuracy,ECL,Time-LLM,MSE,0.270,
+accuracy,Traffic,Time-LLM,MSE,0.429,
+accuracy,Traffic,Time-LLM,MSE,0.306,
diff --git a/result/per_paper/2310.01728/accuracy_efficiency_traced.csv b/result/per_paper/2310.01728/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..0dff4ad6563f49dc7e0283df2dabf23c6eefabe5
--- /dev/null
+++ b/result/per_paper/2310.01728/accuracy_efficiency_traced.csv
@@ -0,0 +1,90 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,Time-LLM,MSE,0.613,,1,2,14
+accuracy,Avg,Time-LLM,MSE,0.596,,1,6,14
+accuracy,ETTh2,Time-LLM,MSE,0.413,,1,7,14
+accuracy,Avg,Time-LLM,MSE,0.499,,1,11,14
+accuracy,ETTm1,Time-LLM,MSE,0.774,,1,12,14
+accuracy,Avg,Time-LLM,MSE,0.628,,1,16,14
+accuracy,ETTm2,Time-LLM,MSE,0.352,,1,17,14
+accuracy,Avg,Time-LLM,MSE,0.930,,1,21,14
+accuracy,Weather,Time-LLM,MSE,0.221,,1,22,14
+accuracy,Avg,Time-LLM,MSE,0.342,,1,26,14
+accuracy,Electricity,Time-LLM,MSE,0.261,,1,27,14
+accuracy,Avg,Time-LLM,MSE,0.478,,1,31,14
+accuracy,Traffic,Time-LLM,MSE,0.672,,1,32,14
+accuracy,Avg,Time-LLM,MSE,0.446,,1,36,14
+accuracy,ETTh1,Time-LLM,MSE,0.681,,2,2,14
+accuracy,Avg,Time-LLM,MSE,0.598,,2,6,14
+accuracy,ETTh2,Time-LLM,MSE,0.428,,2,7,14
+accuracy,Avg,Time-LLM,MSE,0.457,,2,11,14
+accuracy,ETTm1,Time-LLM,MSE,0.726,,2,12,14
+accuracy,Avg,Time-LLM,MSE,0.620,,2,16,14
+accuracy,ETTm2,Time-LLM,MSE,0.232,,2,17,14
+accuracy,Avg,Time-LLM,MSE,0.433,,2,21,14
+accuracy,Weather,Time-LLM,MSE,0.227,,2,22,14
+accuracy,Avg,Time-LLM,MSE,0.353,,2,26,14
+accuracy,Electricity,Time-LLM,MSE,0.297,,2,27,14
+accuracy,Avg,Time-LLM,MSE,0.404,,2,31,14
+accuracy,Traffic,Time-LLM,MSE,0.795,,2,32,14
+accuracy,Avg,Time-LLM,MSE,0.502,,2,36,14
+accuracy,$ETTh1 \rightarrow ETTh2$,Time-LLM,MSE,0.400,,3,2,9
+accuracy,Avg,Time-LLM,MSE,0.380,,3,6,9
+accuracy,$ETTh1 \rightarrow ETTm2$,Time-LLM,MSE,0.357,,3,7,9
+accuracy,Avg,Time-LLM,MSE,0.314,,3,11,9
+accuracy,$ETTh2 \rightarrow ETTh1$,Time-LLM,MSE,0.555,,3,12,9
+accuracy,Avg,Time-LLM,MSE,0.565,,3,16,9
+accuracy,$ETTh2 \rightarrow ETTm2$,Time-LLM,MSE,0.336,,3,17,9
+accuracy,Avg,Time-LLM,MSE,0.325,,3,21,9
+accuracy,$ETTm1 \rightarrow ETTh2$,Time-LLM,MSE,0.415,,3,22,9
+accuracy,Avg,Time-LLM,MSE,0.439,,3,26,9
+accuracy,$ETTm1 \rightarrow ETTm2$,Time-LLM,MSE,0.314,,3,27,9
+accuracy,Avg,Time-LLM,MSE,0.296,,3,31,9
+accuracy,$ETTm2 \rightarrow ETTh2$,Time-LLM,MSE,0.391,,3,32,9
+accuracy,Avg,Time-LLM,MSE,0.409,,3,36,9
+accuracy,$ETTm2 \rightarrow ETTm1$,Time-LLM,MSE,0.490,,3,37,9
+accuracy,Avg,Time-LLM,MSE,0.568,,3,41,9
+accuracy,ETTh1,Time-LLM,MSE,0.933,,4,2,8
+accuracy,Avg,Time-LLM,MSE,0.656,,4,6,8
+accuracy,ETTh2,Time-LLM,MSE,0.390,,4,7,8
+accuracy,Avg,Time-LLM,MSE,0.532,,4,11,8
+accuracy,ETTm1,Time-LLM,MSE,1.091,,4,12,8
+accuracy,Avg,Time-LLM,MSE,0.697,,4,16,8
+accuracy,ETTm2,Time-LLM,MSE,0.435,,4,17,8
+accuracy,Avg,Time-LLM,MSE,0.634,,4,21,8
+accuracy,Weather,Time-LLM,MSE,0.255,,4,22,8
+accuracy,Avg,Time-LLM,MSE,0.448,,4,26,8
+accuracy,Electricity,Time-LLM,MSE,0.520,,4,27,8
+accuracy,Avg,Time-LLM,MSE,0.510,,4,31,8
+accuracy,Traffic,Time-LLM,MSE,1.068,,4,32,8
+accuracy,Avg,Time-LLM,MSE,0.765,,4,36,8
+accuracy,ILI,Time-LLM,MSE,4.909,,4,37,8
+accuracy,Avg,Time-LLM,MSE,1.346,,4,41,8
+accuracy,Yearly,Time-LLM,SMAPE,13.419,,5,1,2
+accuracy,MASE,Time-LLM,SMAPE,3.565,,5,2,2
+accuracy,OWA,Time-LLM,SMAPE,0.911,,5,3,2
+accuracy,Quarterly,Time-LLM,SMAPE,10.110,,5,4,2
+accuracy,MASE,Time-LLM,SMAPE,1.253,,5,5,2
+accuracy,OWA,Time-LLM,SMAPE,0.938,,5,6,2
+accuracy,Monthly,Time-LLM,SMAPE,12.980,,5,7,2
+accuracy,MASE,Time-LLM,SMAPE,1.003,,5,8,2
+accuracy,OWA,Time-LLM,SMAPE,0.931,,5,9,2
+accuracy,Others,Time-LLM,SMAPE,4.795,,5,10,2
+accuracy,MASE,Time-LLM,SMAPE,4.116,,5,11,2
+accuracy,OWA,Time-LLM,SMAPE,1.259,,5,12,2
+accuracy,Average,Time-LLM,SMAPE,11.983,,5,13,2
+accuracy,MASE,Time-LLM,SMAPE,1.808,,5,14,2
+accuracy,OWA,Time-LLM,SMAPE,0.94,,5,15,2
+accuracy,ETTh1,Time-LLM,MSE,0.556,,7,2,1
+accuracy,ETTh1,Time-LLM,MSE,0.522,,7,2,2
+accuracy,ETTh2,Time-LLM,MSE,0.370,,7,3,1
+accuracy,ETTh2,Time-LLM,MSE,0.394,,7,3,2
+accuracy,ETTm1,Time-LLM,MSE,0.404,,7,4,1
+accuracy,ETTm1,Time-LLM,MSE,0.427,,7,4,2
+accuracy,ETTm2,Time-LLM,MSE,0.277,,7,5,1
+accuracy,ETTm2,Time-LLM,MSE,0.323,,7,5,2
+accuracy,Weather,Time-LLM,MSE,0.234,,7,6,1
+accuracy,Weather,Time-LLM,MSE,0.273,,7,6,2
+accuracy,ECL,Time-LLM,MSE,0.175,,7,7,1
+accuracy,ECL,Time-LLM,MSE,0.270,,7,7,2
+accuracy,Traffic,Time-LLM,MSE,0.429,,7,8,1
+accuracy,Traffic,Time-LLM,MSE,0.306,,7,8,2
diff --git a/result/per_paper/2310.01728/components_architecture.csv b/result/per_paper/2310.01728/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2feabc32f3386134f14d8f88b04aaf91f451f897
--- /dev/null
+++ b/result/per_paper/2310.01728/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Text Prototype Representation,A method to reprogram input time series into text-like token sequences that align with the modality of large language models (LLMs).,proposed_here,,We begin by reprogramming the input time series with text prototypes before feeding it into the frozen LLM to align the two modalities.
+Prompt-as-Prefix (PaP),"A technique to enrich input context with declarative prompts (e.g., domain knowledge, task instructions) to guide the LLM's reasoning about time series data.",proposed_here,,"We propose Prompt-as-Prefix (PaP), which enriches the input context and directs the transformation of reprogrammed input patches."
+Forecast Projection Module,A component that projects the output of the LLM (after processing reprogrammed time series) into time series forecasts.,proposed_here,,The transformed time series patches from the LLM are finally projected to obtain the forecasts.
+Frozen LLM Backbone,"The pre-trained large language model (e.g., GPT variants) used as a fixed backbone for processing reprogrammed time series inputs.",reused_cited,"Chen, 2022; Zhou et al., 2023a","Our approach aligns with model reprogramming (Chen, 2022) and leverages pre-trained language models without altering their self-attention and feedforward layers (Zhou et al., 2023a)."
diff --git a/result/per_paper/2310.01728/computational.csv b/result/per_paper/2310.01728/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ae40b71244f0619509c069d5fd8bbcfe0ee3ed71
--- /dev/null
+++ b/result/per_paper/2310.01728/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA A100-80G,GPU,stated,Our experiments are conducted on NVIDIA A100-80G GPUs.
+num_devices,,,not_reported,"Number of devices (e.g., GPUs) is not explicitly stated."
+training_cost,,,not_reported,"Training cost (e.g., monetary or computational cost) is not reported."
+training_batch_size,,,not_reported,"Batch sizes vary per task (e.g., 16 for ETTh1, 8 for Weather), but no single value is stated for Time-LLM."
+training_steps_or_epochs,,,not_reported,"Epochs vary per task (e.g., 50 for ETTh1, 100 for ETTm1), but no single value is stated for Time-LLM."
+precision,,,not_reported,"Precision (e.g., 16-bit/32-bit) is not explicitly stated."
+inference_latency,,,not_reported,Inference latency is not mentioned.
+inference_throughput,,,not_reported,Inference throughput is not mentioned.
+peak_memory,,,not_reported,"Peak memory is not explicitly stated, though GPU memory overhead is mentioned as a metric."
+flops_or_macs,,,not_reported,FLOPs or MACs are not mentioned.
+num_inference_samples,,,not_reported,Number of inference samples is not mentioned.
+params,,,not_reported,"Trainable parameters are mentioned as a metric, but no absolute value is provided for Time-LLM."
+context_lengths_evaluated,,,not_reported,"Input length $T$ is mentioned (e.g., 512 for LTF tasks), but not explicitly stated as 'context lengths evaluated.'"
+horizon_lengths_evaluated,"96, 336",steps,stated,Forecasting tasks on ETTh1 involve horizons of 96 and 336 steps.
+inference_batch_size,,,not_reported,Inference batch size is not mentioned.
diff --git a/result/per_paper/2310.03589/accuracy_efficiency.csv b/result/per_paper/2310.03589/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2310.03589/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2310.03589/accuracy_efficiency_traced.csv b/result/per_paper/2310.03589/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2310.03589/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2310.03589/components_architecture.csv b/result/per_paper/2310.03589/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..35c563a283ebca4a23854943b18bd83c8143218e
--- /dev/null
+++ b/result/per_paper/2310.03589/components_architecture.csv
@@ -0,0 +1,10 @@
+component,what_it_is,provenance,citation,evidence
+Transformer-based Architecture,"The foundational structure of TimeGPT, utilizing self-attention mechanisms to process temporal dependencies.",reused_cited,"Vaswani et al., 2017","TimeGPT is a Transformer-based time series model with self-attention mechanisms based on [Vaswani et al., 2017]."
+Positional Encoding,A method to inject information about the position of elements in the sequence into the model's input.,reused_cited,"Vaswani et al., 2017","The architecture consists of an encoder-decoder structure with multiple layers, each with residual connections and layer normalization. Finally, a linear layer maps the decoder's output to the forecasting window dimension."
+Encoder-Decoder Structure,A two-part architecture where the encoder processes input sequences and the decoder generates output sequences.,reused_cited,"Vaswani et al., 2017","The architecture consists of an encoder-decoder structure with multiple layers, each with residual connections and layer normalization."
+Residual Connections,A technique to facilitate the training of deeper networks by allowing gradients to flow through the network more effectively.,reused_cited,"Vaswani et al., 2017","The architecture consists of an encoder-decoder structure with multiple layers, each with residual connections and layer normalization."
+Layer Normalization,A normalization technique applied to the outputs of each layer to stabilize the training process.,reused_cited,"Vaswani et al., 2017","The architecture consists of an encoder-decoder structure with multiple layers, each with residual connections and layer normalization."
+Multi-head Attention,A mechanism that allows the model to focus on different parts of the input sequence simultaneously.,reused_cited,"Vaswani et al., 2017",The general intuition is that attention-based mechanisms are able to capture the diversity of past events and correctly extrapolate potential future distributions.
+CNN (Convolutional Neural Network),A component integrated into the model to process local patterns in the time series data.,proposed_here,,"The architecture includes a CNN as part of the processing pipeline, as indicated in the mermaid diagram."
+Linear Layer,A final layer that maps the decoder's output to the forecasting window dimension.,proposed_here,,"Finally, a linear layer maps the decoder's output to the forecasting window dimension."
+Output Embedding,A representation of the output sequence used during both training and inference phases.,proposed_here,,The mermaid diagram shows 'Output Embedding' as a component in both training and inference subgraphs.
diff --git a/result/per_paper/2310.03589/computational.csv b/result/per_paper/2310.03589/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..709163c92148a91f3be273bf116d5071baba2d3c
--- /dev/null
+++ b/result/per_paper/2310.03589/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA A10G GPUs,,stated,TimeGPT underwent a multi-day training period on a cluster of NVIDIA A10G GPUs.
+num_devices,,,not_reported,The paper does not specify the number of devices in the cluster.
+training_cost,,,not_reported,"The paper does not report training cost (e.g., monetary or energy cost)."
+training_batch_size,,,not_reported,The paper mentions hyperparameter exploration but does not specify the training batch size.
+training_steps_or_epochs,,,not_reported,The paper does not report training steps or epochs.
+precision,,,not_reported,"The paper does not specify training or inference precision (e.g., FP16, FP32)."
+inference_latency,0.6,milliseconds per series,stated,"For zero-shot inference, our internal tests recorded an average GPU inference speed of 0.6 milliseconds per series for TimeGPT."
+inference_throughput,,,not_reported,"The paper does not report inference throughput (e.g., samples per second)."
+peak_memory,,,not_reported,The paper does not report peak memory usage during training or inference.
+flops_or_macs,,,not_reported,The paper does not report FLOPs or MACs for the model.
+num_inference_samples,,,not_reported,The paper does not specify the number of inference samples tested.
+params,,,not_reported,The paper does not report the number of parameters in TimeGPT-1.
+context_lengths_evaluated,,,not_reported,The paper does not specify context lengths evaluated during training or testing.
+horizon_lengths_evaluated,,,not_reported,The paper does not specify forecast horizon lengths evaluated.
+inference_batch_size,,,not_reported,The paper does not report inference batch size.
diff --git a/result/per_paper/2310.04948/accuracy_efficiency.csv b/result/per_paper/2310.04948/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2b1e8314c7f34af087ff64717e11ee1bc25b20cf
--- /dev/null
+++ b/result/per_paper/2310.04948/accuracy_efficiency.csv
@@ -0,0 +1,63 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,GPT2,TEMPO,MSE/MAE,0.193,
+accuracy,T5,TEMPO,MSE/MAE,0.185,
+accuracy,PatchTST,TEMPO,MSE/MAE,0.489,
+accuracy,Timesnet,TEMPO,MSE/MAE,0.293,
+accuracy,FEDformer,TEMPO,MSE/MAE,0.300,
+accuracy,ETSformer,TEMPO,MSE/MAE,0.707,
+accuracy,Informer,TEMPO,MSE/MAE,0.512,
+accuracy,DLinear,TEMPO,MSE/MAE,0.195,
+accuracy,GPT2,TEMPO,MSE/MAE,0.207,
+accuracy,T5,TEMPO,MSE/MAE,0.205,
+accuracy,PatchTST,TEMPO,MSE/MAE,0.465,
+accuracy,Timesnet,TEMPO,MSE/MAE,0.283,
+accuracy,FEDformer,TEMPO,MSE/MAE,0.390,
+accuracy,ETSformer,TEMPO,MSE/MAE,0.721,
+accuracy,Informer,TEMPO,MSE/MAE,0.625,
+accuracy,DLinear,TEMPO,MSE/MAE,0.204,
+accuracy,GPT2,TEMPO,MSE/MAE,0.231,
+accuracy,T5,TEMPO,MSE/MAE,0.229,
+accuracy,PatchTST,TEMPO,MSE/MAE,0.531,
+accuracy,Timesnet,TEMPO,MSE/MAE,0.733,
+accuracy,FEDformer,TEMPO,MSE/MAE,0.317,
+accuracy,ETSformer,TEMPO,MSE/MAE,0.862,
+accuracy,Informer,TEMPO,MSE/MAE,1.222,
+accuracy,DLinear,TEMPO,MSE/MAE,0.231,
+accuracy,GPT2,TEMPO,MSE/MAE,0.262,
+accuracy,T5,TEMPO,MSE/MAE,0.266,
+accuracy,PatchTST,TEMPO,MSE/MAE,0.475,
+accuracy,Timesnet,TEMPO,MSE/MAE,1.166,
+accuracy,FEDformer,TEMPO,MSE/MAE,0.423,
+accuracy,ETSformer,TEMPO,MSE/MAE,0.666,
+accuracy,Informer,TEMPO,MSE/MAE,0.881,
+accuracy,DLinear,TEMPO,MSE/MAE,0.259,
+accuracy,GPT2,TEMPO,MSE/MAE,0.223,
+accuracy,T5,TEMPO,MSE/MAE,0.221,
+accuracy,PatchTST,TEMPO,MSE/MAE,0.49,
+accuracy,Timesnet,TEMPO,MSE/MAE,0.619,
+accuracy,FEDformer,TEMPO,MSE/MAE,0.358,
+accuracy,ETSformer,TEMPO,MSE/MAE,0.750,
+accuracy,Informer,TEMPO,MSE/MAE,0.810,
+accuracy,DLinear,TEMPO,MSE/MAE,0.222,
+accuracy,ETTh1,TEMPO,MSE/MAE,0.400,
+accuracy,Avg.,TEMPO,MSE/MAE,0.454,
+accuracy,ETTh2,TEMPO,MSE/MAE,0.301,
+accuracy,Avg.,TEMPO,MSE/MAE,0.383,
+accuracy,ECL,TEMPO,MSE/MAE,0.19,
+accuracy,Avg,TEMPO,MSE/MAE,0.277,
+accuracy,Traffic,TEMPO,MSE/MAE,0.56,
+accuracy,Avg,TEMPO,MSE/MAE,0.621,
+accuracy,Weather,TEMPO,MSE/MAE,0.217,
+accuracy,Avg,TEMPO,MSE/MAE,0.328,
+accuracy,CC,TEMPO,SMAPE,32.27,
+accuracy,CD,TEMPO,SMAPE,25.9,
+accuracy,Ind,TEMPO,SMAPE,26.7,
+accuracy,RE,TEMPO,SMAPE,29.46,
+accuracy,ECL,TEMPO,MSE/MAE,0.178,
+accuracy,Avg,TEMPO,MSE/MAE,0.228,
+accuracy,Ettm1,TEMPO,MSE/MAE,0.438,
+accuracy,Avg,TEMPO,MSE/MAE,0.575,
+accuracy,Avg.,TEMPO,MSE/MAE,0.216,
+accuracy,Avg.,TEMPO,MSE/MAE,0.503,
+accuracy,Avg.,TEMPO,MSE/MAE,0.287,
+accuracy,Avg.,TEMPO,MSE/MAE,0.280,
diff --git a/result/per_paper/2310.04948/accuracy_efficiency_traced.csv b/result/per_paper/2310.04948/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..292880948cd64b4a2a955dda17e946e714292ff3
--- /dev/null
+++ b/result/per_paper/2310.04948/accuracy_efficiency_traced.csv
@@ -0,0 +1,63 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,GPT2,TEMPO,MSE/MAE,0.193,,0,3,1
+accuracy,T5,TEMPO,MSE/MAE,0.185,,0,4,1
+accuracy,PatchTST,TEMPO,MSE/MAE,0.489,,0,5,1
+accuracy,Timesnet,TEMPO,MSE/MAE,0.293,,0,6,1
+accuracy,FEDformer,TEMPO,MSE/MAE,0.300,,0,7,1
+accuracy,ETSformer,TEMPO,MSE/MAE,0.707,,0,8,1
+accuracy,Informer,TEMPO,MSE/MAE,0.512,,0,9,1
+accuracy,DLinear,TEMPO,MSE/MAE,0.195,,0,10,1
+accuracy,GPT2,TEMPO,MSE/MAE,0.207,,0,12,1
+accuracy,T5,TEMPO,MSE/MAE,0.205,,0,13,1
+accuracy,PatchTST,TEMPO,MSE/MAE,0.465,,0,14,1
+accuracy,Timesnet,TEMPO,MSE/MAE,0.283,,0,15,1
+accuracy,FEDformer,TEMPO,MSE/MAE,0.390,,0,16,1
+accuracy,ETSformer,TEMPO,MSE/MAE,0.721,,0,17,1
+accuracy,Informer,TEMPO,MSE/MAE,0.625,,0,18,1
+accuracy,DLinear,TEMPO,MSE/MAE,0.204,,0,19,1
+accuracy,GPT2,TEMPO,MSE/MAE,0.231,,0,21,1
+accuracy,T5,TEMPO,MSE/MAE,0.229,,0,22,1
+accuracy,PatchTST,TEMPO,MSE/MAE,0.531,,0,23,1
+accuracy,Timesnet,TEMPO,MSE/MAE,0.733,,0,24,1
+accuracy,FEDformer,TEMPO,MSE/MAE,0.317,,0,25,1
+accuracy,ETSformer,TEMPO,MSE/MAE,0.862,,0,26,1
+accuracy,Informer,TEMPO,MSE/MAE,1.222,,0,27,1
+accuracy,DLinear,TEMPO,MSE/MAE,0.231,,0,28,1
+accuracy,GPT2,TEMPO,MSE/MAE,0.262,,0,30,1
+accuracy,T5,TEMPO,MSE/MAE,0.266,,0,31,1
+accuracy,PatchTST,TEMPO,MSE/MAE,0.475,,0,32,1
+accuracy,Timesnet,TEMPO,MSE/MAE,1.166,,0,33,1
+accuracy,FEDformer,TEMPO,MSE/MAE,0.423,,0,34,1
+accuracy,ETSformer,TEMPO,MSE/MAE,0.666,,0,35,1
+accuracy,Informer,TEMPO,MSE/MAE,0.881,,0,36,1
+accuracy,DLinear,TEMPO,MSE/MAE,0.259,,0,37,1
+accuracy,GPT2,TEMPO,MSE/MAE,0.223,,0,39,1
+accuracy,T5,TEMPO,MSE/MAE,0.221,,0,40,1
+accuracy,PatchTST,TEMPO,MSE/MAE,0.49,,0,41,1
+accuracy,Timesnet,TEMPO,MSE/MAE,0.619,,0,42,1
+accuracy,FEDformer,TEMPO,MSE/MAE,0.358,,0,43,1
+accuracy,ETSformer,TEMPO,MSE/MAE,0.750,,0,44,1
+accuracy,Informer,TEMPO,MSE/MAE,0.810,,0,45,1
+accuracy,DLinear,TEMPO,MSE/MAE,0.222,,0,46,1
+accuracy,ETTh1,TEMPO,MSE/MAE,0.400,,1,2,2
+accuracy,Avg.,TEMPO,MSE/MAE,0.454,,1,6,2
+accuracy,ETTh2,TEMPO,MSE/MAE,0.301,,1,7,2
+accuracy,Avg.,TEMPO,MSE/MAE,0.383,,1,11,2
+accuracy,ECL,TEMPO,MSE/MAE,0.19,,3,2,2
+accuracy,Avg,TEMPO,MSE/MAE,0.277,,3,6,2
+accuracy,Traffic,TEMPO,MSE/MAE,0.56,,3,7,2
+accuracy,Avg,TEMPO,MSE/MAE,0.621,,3,11,2
+accuracy,Weather,TEMPO,MSE/MAE,0.217,,3,12,2
+accuracy,Avg,TEMPO,MSE/MAE,0.328,,3,16,2
+accuracy,CC,TEMPO,SMAPE,32.27,,4,2,1
+accuracy,CD,TEMPO,SMAPE,25.9,,4,3,1
+accuracy,Ind,TEMPO,SMAPE,26.7,,4,4,1
+accuracy,RE,TEMPO,SMAPE,29.46,,4,5,1
+accuracy,ECL,TEMPO,MSE/MAE,0.178,,5,2,2
+accuracy,Avg,TEMPO,MSE/MAE,0.228,,5,6,2
+accuracy,Ettm1,TEMPO,MSE/MAE,0.438,,5,7,2
+accuracy,Avg,TEMPO,MSE/MAE,0.575,,5,11,2
+accuracy,Avg.,TEMPO,MSE/MAE,0.216,,6,7,1
+accuracy,Avg.,TEMPO,MSE/MAE,0.503,,6,7,3
+accuracy,Avg.,TEMPO,MSE/MAE,0.287,,6,7,5
+accuracy,Avg.,TEMPO,MSE/MAE,0.280,,6,7,7
diff --git a/result/per_paper/2310.04948/components_architecture.csv b/result/per_paper/2310.04948/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d9c67d7d74a84d31ec53f42d32b8a773c573ee5b
--- /dev/null
+++ b/result/per_paper/2310.04948/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Time Series Decomposition Module,"Decomposes input time series into three additive components: trend, seasonality, and residuals using locally weighted scatterplot smoothing (LOESS).",proposed_here,,"TEMPO first decomposes time series input into three additive components, i.e., trend, seasonality, and residuals via locally weighted scatterplot smoothing (Cleveland et al., 1990)."
+Hidden Space Mapping,"Maps each decomposed time series component (trend, seasonality, residuals) to a corresponding hidden space to construct the input embedding for the generative pre-trained transformer (GPT).",proposed_here,,Each of these temporal inputs is subsequently mapped to its corresponding hidden space to construct the time series input embedding of the generative pre-trained transformer (GPT).
+Soft Prompt Mechanism,"A prompt-based tuning strategy that uses learnable continuous vector representations to encode temporal knowledge of trend and seasonality, guiding the GPT for forecasting tasks.",proposed_here,,TEMPO utilizes a soft prompt to efficiently tune the GPT [...] by guiding the reuse of a collection of learnable continuous vector representations that encode temporal knowledge of trend and seasonality.
+Generative Pre-trained Transformer (GPT) Core,"The foundational transformer architecture adapted for time series forecasting, leveraging pre-trained weights and attention mechanisms.",reused_cited,"Radford, A., et al. (2019)",TEMPO [...] expands the capability for dynamically modeling real-world temporal phenomena from data within diverse domains.
+Interpretable Framework for Component Interaction,"A framework that leverages the three additive components (trend, seasonality, residuals) to provide interpretable insights into interactions among input components.",proposed_here,"Hastie, T. (2017)","We leverage the three key additive components of time series data [...] to provide an interpretable framework for comprehending the interactions among input components (Hastie, 2017)."
diff --git a/result/per_paper/2310.04948/computational.csv b/result/per_paper/2310.04948/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2310.04948/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2310.07446/accuracy_efficiency.csv b/result/per_paper/2310.07446/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2310.07446/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2310.07446/accuracy_efficiency_traced.csv b/result/per_paper/2310.07446/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2310.07446/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2310.07446/components_architecture.csv b/result/per_paper/2310.07446/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..0cb4f29aa1e63173b207634719a547c26d95619e
--- /dev/null
+++ b/result/per_paper/2310.07446/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Distributional Forecasting Approach,"Method for generating forecasts, ranging from point forecasts to predefined distribution heads based on data assumptions.",reused_cited,"56, 70","The first aspect involves the approach to distributional forecasting, ranging from models focused on point forecasts [49, 40] to those using pre-defined distribution heads based on specific data assumptions [56, 70]."
+Decoding Scheme,"Strategy for generating multi-step forecasts, either autoregressive (AR) or non-autoregressive (NAR).",reused_cited,"58, 57, 32","The second aspect is the decoding scheme used to generate multi-step forecasts, which can be either autoregressive (AR) or non-autoregressive (NAR)."
+Normalization Strategy,"Choice of normalization technique, such as reversible instance normalization (RevIN) for long-term point forecasting or mean scaling for short-term probabilistic forecasting.",reused_cited,"32, 58, 57","The third aspect pertains to the normalization choice, where the long-term point forecasting models typically employ reversible instance normalization (RevIN) [32] while short-term probabilistic ones often use mean scaling strategies [58, 57]."
+Trend and Seasonality Strength Measurement,Metric to quantify the strength of trends and seasonality in time-series data for benchmarking.,proposed_here,,"We measure three essential data characteristics in ProbTS: the strength of trends and seasonality, and the complexity of the data distribution."
+Data Distribution Complexity Measurement,Metric to assess the complexity of data distributions in time-series for benchmarking.,proposed_here,,"We measure three essential data characteristics in ProbTS: the strength of trends and seasonality, and the complexity of the data distribution."
diff --git a/result/per_paper/2310.07446/computational.csv b/result/per_paper/2310.07446/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..7d26c296b482d0998b7e84ded3a7b9d3f2c667b5
--- /dev/null
+++ b/result/per_paper/2310.07446/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA Tesla V100,GPU,stated,run on single NVIDIA Tesla V100 GPUs with CUDA 11.3
+num_devices,1,device,stated,run on single NVIDIA Tesla V100 GPUs
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,50,epochs,stated,limited training to 50 epochs
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,100,samples,stated,100 samples are employed to estimate the empirical CDF
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,"[24, 48, 96, 192, 336, 720]",time steps,stated,"selected prediction horizons of {24, 48, 96, 192, 336, 720}"
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2310.08278/accuracy_efficiency.csv b/result/per_paper/2310.08278/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..c33055058fccd7168fa54f2075be1c28e0a239d0
--- /dev/null
+++ b/result/per_paper/2310.08278/accuracy_efficiency.csv
@@ -0,0 +1,13 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,PATCHTST,Lag-Llama,CRPS,0.169,
+accuracy,TFT,Lag-Llama,CRPS,0.154,
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.136,
+accuracy,PATCHTST,Lag-Llama,CRPS,0.171,
+accuracy,TFT,Lag-Llama,CRPS,0.156,
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.135,
+accuracy,PATCHTST,Lag-Llama,CRPS,0.174,
+accuracy,TFT,Lag-Llama,CRPS,0.152,
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.133,
+accuracy,PATCHTST,Lag-Llama,CRPS,0.174,
+accuracy,TFT,Lag-Llama,CRPS,0.148,
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.132,
diff --git a/result/per_paper/2310.08278/accuracy_efficiency_traced.csv b/result/per_paper/2310.08278/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..caac50a70461211413090e68a351aa133ff42bfc
--- /dev/null
+++ b/result/per_paper/2310.08278/accuracy_efficiency_traced.csv
@@ -0,0 +1,13 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,PATCHTST,Lag-Llama,CRPS,0.169,,0,3,1
+accuracy,TFT,Lag-Llama,CRPS,0.154,,0,4,1
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.136,,0,5,1
+accuracy,PATCHTST,Lag-Llama,CRPS,0.171,,0,7,1
+accuracy,TFT,Lag-Llama,CRPS,0.156,,0,8,1
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.135,,0,9,1
+accuracy,PATCHTST,Lag-Llama,CRPS,0.174,,0,11,1
+accuracy,TFT,Lag-Llama,CRPS,0.152,,0,12,1
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.133,,0,13,1
+accuracy,PATCHTST,Lag-Llama,CRPS,0.174,,0,15,1
+accuracy,TFT,Lag-Llama,CRPS,0.148,,0,16,1
+accuracy,LAG-LLAMA,Lag-Llama,CRPS,0.132,,0,17,1
diff --git a/result/per_paper/2310.08278/components_architecture.csv b/result/per_paper/2310.08278/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b3d5f23e56abe30b1a140e9149baceb2d125b7ae
--- /dev/null
+++ b/result/per_paper/2310.08278/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Decoder-only Transformer Architecture,"A transformer-based model structure that uses only the decoder component, designed for sequence generation tasks.",proposed_here,"Rasul et al., 2023",Lag-Llama is based on a decoder-only transformer architecture that uses lags as covariates.
+Lags as Covariates,A method where historical time series values (lags) are used as input features to predict future values.,proposed_here,"Rasul et al., 2023",uses lags as covariates.
+Pretraining on Diverse Time Series Data,"A pretraining strategy involving a large, diverse corpus of time series data from multiple domains to enable zero-shot generalization.",proposed_here,"Rasul et al., 2023",pretrained on a large corpus of diverse time series data from several domains.
+Probabilistic Forecasting Mechanism,"A framework for generating probabilistic forecasts, providing uncertainty estimates alongside point predictions.",proposed_here,"Rasul et al., 2023",univariate probabilistic time series forecasting.
+Univariate Time Series Processing,A focus on modeling single-variable (univariate) time series data rather than multivariate sequences.,proposed_here,"Rasul et al., 2023",univariate probabilistic time series forecasting.
diff --git a/result/per_paper/2310.08278/computational.csv b/result/per_paper/2310.08278/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..fd1379d2677a38bf1d77df47f4ffbcb89186aa5d
--- /dev/null
+++ b/result/per_paper/2310.08278/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,Nvidia Tesla-P100 GPU,,stated,"We use a single Nvidia Tesla-P100 GPU with 12 GB of memory, 4 CPU cores, and 24 GB of RAM."
+num_devices,1,,stated,We use a single Nvidia Tesla-P100 GPU...
+training_cost,,,not_reported,
+training_batch_size,256,,stated,"During pretraining, we use the batch size of 256..."
+training_steps_or_epochs,50,epochs,stated,We use an early stopping criterion of 50 epochs based on the average validation loss...
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2310.10688/accuracy_efficiency.csv b/result/per_paper/2310.10688/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..32adb50b13e1d775d53dfd97333a1a2c4ca8cff3
--- /dev/null
+++ b/result/per_paper/2310.10688/accuracy_efficiency.csv
@@ -0,0 +1,48 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,australian electricity demand,TimesFM(ZS),MAE,448.81,
+accuracy,bitcoin,TimesFM(ZS),MAE,1.3,
+accuracy,pedestrian counts,TimesFM(ZS),MAE,40.71,
+accuracy,weather,TimesFM(ZS),MAE,2.07,
+accuracy,nn5 daily,TimesFM(ZS),MAE,3.54,
+accuracy,nn5 weekly,TimesFM(ZS),MAE,14.67,
+accuracy,tourism yearly,TimesFM(ZS),MAE,109977.29,
+accuracy,tourism quarterly,TimesFM(ZS),MAE,12102.04,
+accuracy,tourism monthly,TimesFM(ZS),MAE,3183.77,
+accuracy,cif 2016,TimesFM(ZS),MAE,773980.44,
+accuracy,covid deaths,TimesFM(ZS),MAE,209.80,
+accuracy,fred md,TimesFM(ZS),MAE,947.12,
+accuracy,traffic hourly,TimesFM(ZS),MAE,0.01,
+accuracy,traffic weekly,TimesFM(ZS),MAE,1.12,
+accuracy,saugeenday,TimesFM(ZS),MAE,24.63,
+accuracy,us births,TimesFM(ZS),MAE,437.27,
+accuracy,hospital,TimesFM(ZS),MAE,19.41,
+accuracy,solar weekly,TimesFM(ZS),MAE,1258.27,
+accuracy,Scaled MAE (Arithmetic Mean),TimesFM(ZS),MAE,0.8005,
+accuracy,Scaled MAE (Geometric Mean),TimesFM(ZS),MAE,0.6846,
+accuracy,ETTh1,TimesFM,MAE,96,
+accuracy,Avg,TimesFM,MAE,0.426,
+accuracy,ETTh2,TimesFM,MAE,96,
+accuracy,Avg,TimesFM,MAE,0.410,
+accuracy,ETTm1,TimesFM,MAE,96,
+accuracy,Avg,TimesFM,MAE,0.388,
+accuracy,ETTm2,TimesFM,MAE,96,
+accuracy,Avg,TimesFM,MAE,0.334,
+accuracy,AirPassengersDataset,TimesFM(ZS),MAE,62.51,
+accuracy,AusBeerDataset,TimesFM(ZS),MAE,11.94,
+accuracy,GasRateCO2Dataset,TimesFM(ZS),MAE,2.50,
+accuracy,MonthlyMilkDataset,TimesFM(ZS),MAE,28.09,
+accuracy,SunspotsDataset,TimesFM(ZS),MAE,41.40,
+accuracy,WineDataset,TimesFM(ZS),MAE,2871.33,
+accuracy,WoolyDataset,TimesFM(ZS),MAE,728.92,
+accuracy,HeartRateDataset,TimesFM(ZS),MAE,5.85,
+accuracy,Scaled MAE (Arithmetic Mean),TimesFM(ZS),MAE,0.6829,
+accuracy,Scaled MAE (Geometric Mean),TimesFM(ZS),MAE,0.5767,
+accuracy,ETTh1 (horizon=96),TimesFM(ZS),MAE,0.45,
+accuracy,ETTh1 (horizon=192),TimesFM(ZS),MAE,0.53,
+accuracy,ETTh2 (horizon=96),TimesFM(ZS),MAE,0.35,
+accuracy,ETTh2 (horizon=192),TimesFM(ZS),MAE,0.62,
+accuracy,ETTm1 (horizon=96),TimesFM(ZS),MAE,0.19,
+accuracy,ETTm1 (horizon=192),TimesFM(ZS),MAE,0.26,
+accuracy,ETTm2 (horizon=96),TimesFM(ZS),MAE,0.24,
+accuracy,ETTm2 (horizon=192),TimesFM(ZS),MAE,0.27,
+accuracy,Avg,TimesFM(ZS),MAE,0.36,
diff --git a/result/per_paper/2310.10688/accuracy_efficiency_traced.csv b/result/per_paper/2310.10688/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..c448861b1682bf16547111339d91f748b3cbe8cf
--- /dev/null
+++ b/result/per_paper/2310.10688/accuracy_efficiency_traced.csv
@@ -0,0 +1,48 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,australian electricity demand,TimesFM(ZS),MAE,448.81,,0,1,15
+accuracy,bitcoin,TimesFM(ZS),MAE,1.3,,0,2,15
+accuracy,pedestrian counts,TimesFM(ZS),MAE,40.71,,0,3,15
+accuracy,weather,TimesFM(ZS),MAE,2.07,,0,4,15
+accuracy,nn5 daily,TimesFM(ZS),MAE,3.54,,0,5,15
+accuracy,nn5 weekly,TimesFM(ZS),MAE,14.67,,0,6,15
+accuracy,tourism yearly,TimesFM(ZS),MAE,109977.29,,0,7,15
+accuracy,tourism quarterly,TimesFM(ZS),MAE,12102.04,,0,8,15
+accuracy,tourism monthly,TimesFM(ZS),MAE,3183.77,,0,9,15
+accuracy,cif 2016,TimesFM(ZS),MAE,773980.44,,0,10,15
+accuracy,covid deaths,TimesFM(ZS),MAE,209.80,,0,11,15
+accuracy,fred md,TimesFM(ZS),MAE,947.12,,0,12,15
+accuracy,traffic hourly,TimesFM(ZS),MAE,0.01,,0,13,15
+accuracy,traffic weekly,TimesFM(ZS),MAE,1.12,,0,14,15
+accuracy,saugeenday,TimesFM(ZS),MAE,24.63,,0,15,15
+accuracy,us births,TimesFM(ZS),MAE,437.27,,0,16,15
+accuracy,hospital,TimesFM(ZS),MAE,19.41,,0,17,15
+accuracy,solar weekly,TimesFM(ZS),MAE,1258.27,,0,18,15
+accuracy,Scaled MAE (Arithmetic Mean),TimesFM(ZS),MAE,0.8005,,0,19,15
+accuracy,Scaled MAE (Geometric Mean),TimesFM(ZS),MAE,0.6846,,0,20,15
+accuracy,ETTh1,TimesFM,MAE,96,,1,1,1
+accuracy,Avg,TimesFM,MAE,0.426,,1,5,1
+accuracy,ETTh2,TimesFM,MAE,96,,1,6,1
+accuracy,Avg,TimesFM,MAE,0.410,,1,10,1
+accuracy,ETTm1,TimesFM,MAE,96,,1,11,1
+accuracy,Avg,TimesFM,MAE,0.388,,1,15,1
+accuracy,ETTm2,TimesFM,MAE,96,,1,16,1
+accuracy,Avg,TimesFM,MAE,0.334,,1,20,1
+accuracy,AirPassengersDataset,TimesFM(ZS),MAE,62.51,,2,1,7
+accuracy,AusBeerDataset,TimesFM(ZS),MAE,11.94,,2,2,7
+accuracy,GasRateCO2Dataset,TimesFM(ZS),MAE,2.50,,2,3,7
+accuracy,MonthlyMilkDataset,TimesFM(ZS),MAE,28.09,,2,4,7
+accuracy,SunspotsDataset,TimesFM(ZS),MAE,41.40,,2,5,7
+accuracy,WineDataset,TimesFM(ZS),MAE,2871.33,,2,6,7
+accuracy,WoolyDataset,TimesFM(ZS),MAE,728.92,,2,7,7
+accuracy,HeartRateDataset,TimesFM(ZS),MAE,5.85,,2,8,7
+accuracy,Scaled MAE (Arithmetic Mean),TimesFM(ZS),MAE,0.6829,,2,9,7
+accuracy,Scaled MAE (Geometric Mean),TimesFM(ZS),MAE,0.5767,,2,10,7
+accuracy,ETTh1 (horizon=96),TimesFM(ZS),MAE,0.45,,3,1,7
+accuracy,ETTh1 (horizon=192),TimesFM(ZS),MAE,0.53,,3,2,7
+accuracy,ETTh2 (horizon=96),TimesFM(ZS),MAE,0.35,,3,3,7
+accuracy,ETTh2 (horizon=192),TimesFM(ZS),MAE,0.62,,3,4,7
+accuracy,ETTm1 (horizon=96),TimesFM(ZS),MAE,0.19,,3,5,7
+accuracy,ETTm1 (horizon=192),TimesFM(ZS),MAE,0.26,,3,6,7
+accuracy,ETTm2 (horizon=96),TimesFM(ZS),MAE,0.24,,3,7,7
+accuracy,ETTm2 (horizon=192),TimesFM(ZS),MAE,0.27,,3,8,7
+accuracy,Avg,TimesFM(ZS),MAE,0.36,,3,9,7
diff --git a/result/per_paper/2310.10688/components_architecture.csv b/result/per_paper/2310.10688/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d8644ee9c4cce2a5cdf1386d14d0531dffacd31b
--- /dev/null
+++ b/result/per_paper/2310.10688/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Decoder-only Attention Architecture,"A transformer-based model trained exclusively in decoder mode, where the model predicts the next patch based on all past patches in parallel across the context window.",proposed_here,,"Our model is trained in decoder-only mode [LSP+18]. In other words, given a sequence of input patches, the model is optimized to predict the next patch as a function of all past patches."
+Input Patching,"A method to divide time-series data into fixed-length segments (patches) during training, analogous to tokenization in language models.",reused_cited,NNSK22,Inspired by the success of patch based modeling in the recent long horizon forecasting work [NNSK22] we also choose to break down the time-series into patches during training.
+Longer Output Patches,"An architectural choice where output patches for prediction are longer than input patches, enabling simultaneous forecasting of multiple future time-steps.",proposed_here,,"We propose a middle ground by allowing our output patches for prediction to be longer than the input patches. As an example, suppose the input patch length is 32 and output patch length is 128."
+Large-Scale Time-Series Corpus,"A pretraining dataset combining real-world (e.g., web search queries, Wikipedia page visits) and synthetic time-series data to ensure diversity and volume for training.",proposed_here,,"Our model is based on pretraining a decoder style attention model with input patching, using a large time-series corpus comprising both real-world and synthetic datasets."
+Zero-shot Forecasting Capability,"The ability to generate accurate forecasts on previously unseen datasets without task-specific fine-tuning, leveraging the pretraining on diverse data.",proposed_here,,"Our model can yield accurate zero-shot forecasts across different domains, forecasting horizons and temporal granularities."
diff --git a/result/per_paper/2310.10688/computational.csv b/result/per_paper/2310.10688/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2310.10688/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2311.01933/accuracy_efficiency.csv b/result/per_paper/2311.01933/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f7f9219ef976c67f091c37a27150a96e0ea9fd9f
--- /dev/null
+++ b/result/per_paper/2311.01933/accuracy_efficiency.csv
@@ -0,0 +1,5 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Data Budget = 50,ForecastPFN,MAE,50,
+accuracy,Data Budget = 500,ForecastPFN,MAE,500,
+accuracy,"ForecastPFN, low noise","ForecastPFN, lowest noise",MSE,0.359,
+accuracy,ForecastPFN,"ForecastPFN, lowest noise",MSE,0.207,
diff --git a/result/per_paper/2311.01933/accuracy_efficiency_traced.csv b/result/per_paper/2311.01933/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e5c8aecc7acaacd9bd49ef4613d055da94f41965
--- /dev/null
+++ b/result/per_paper/2311.01933/accuracy_efficiency_traced.csv
@@ -0,0 +1,5 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Data Budget = 50,ForecastPFN,MAE,50,,0,1,0
+accuracy,Data Budget = 500,ForecastPFN,MAE,500,,0,12,0
+accuracy,"ForecastPFN, low noise","ForecastPFN, lowest noise",MSE,0.359,,2,2,1
+accuracy,ForecastPFN,"ForecastPFN, lowest noise",MSE,0.207,,2,3,1
diff --git a/result/per_paper/2311.01933/components_architecture.csv b/result/per_paper/2311.01933/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..c27cfd8b916d7ded8e03bc5d23c81e3aad616816
--- /dev/null
+++ b/result/per_paper/2311.01933/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Transformer Encoder,"The base architecture of ForecastPFN, consisting of one multi-head attention layer and two feedforward layers.",reused_cited,"2,25,39","As with the original PFNs [2,25,39], we use a transformer [56] as the base architecture. We use an encoder-based transformer, consisting of one multi-head attention layer and two feedforward layer"
+Multi-Head Attention Layer,A component of the transformer encoder that enables the model to attend to different parts of the input sequence simultaneously.,reused_cited,56,"As with the original PFNs [2,25,39], we use a transformer [56] as the base architecture. We use an encoder-based transformer, consisting of one multi-head attention layer and two feedforward layer"
+Feedforward Layers,Two sequential feedforward networks within the transformer encoder that process information after attention mechanisms.,reused_cited,56,"As with the original PFNs [2,25,39], we use a transformer [56] as the base architecture. We use an encoder-based transformer, consisting of one multi-head attention layer and two feedforward layer"
+Multiplicative Noise,A synthetic data generation technique used to balance signal-to-noise ratios across time series by scaling noise relative to the base series.,proposed_here,,We choose multiplicative noise to better balance the amount of signal to noise across all series.
+Weibull Distribution,"A synthetic data distribution parameterized to interpolate between Gaussian and exponential distributions, used to model noise in the synthetic training data.",proposed_here,,"The Weibull distribution is a simple and natural method to parameterize between Gaussian and exponential distributions, two types of distributions that frequently come up in real-world time series."
diff --git a/result/per_paper/2311.01933/computational.csv b/result/per_paper/2311.01933/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6345f7ba712032cdfec534c6bb2be7edd370a151
--- /dev/null
+++ b/result/per_paper/2311.01933/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware type (e.g., GPU/TPU) is explicitly mentioned in the text."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not reported."
+training_cost,,,not_reported,"Training cost (e.g., monetary or computational resources) is not quantified."
+training_batch_size,,,not_reported,The training batch size is not specified.
+training_steps_or_epochs,,,not_reported,The number of training steps or epochs is not reported.
+precision,,,not_reported,"The precision (e.g., 32-bit vs. 16-bit) used during training is not mentioned."
+inference_latency,0.2,seconds,stated,"The text explicitly states: 'producing predictions on a brand new dataset in just a single forward pass, taking 0.2 seconds.'"
+inference_throughput,,,not_reported,"Inference throughput (e.g., samples per second) is not quantified."
+peak_memory,,,not_reported,Peak memory usage during training or inference is not reported.
+flops_or_macs,,,not_reported,FLOPs or MACs (computational complexity) are not mentioned.
+num_inference_samples,,,not_reported,The number of inference samples evaluated is not specified.
+params,,,not_reported,The total number of model parameters is not reported.
+context_lengths_evaluated,,,not_reported,"The context lengths (e.g., input sequence lengths) evaluated during testing are not explicitly stated."
+horizon_lengths_evaluated,,,not_reported,"The prediction horizons (e.g., output sequence lengths) evaluated during testing are not explicitly stated."
+inference_batch_size,,,not_reported,The inference batch size is not specified.
diff --git a/result/per_paper/2311.11413/accuracy_efficiency.csv b/result/per_paper/2311.11413/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..1472a74722e3d14bd7ad8c870fed55939eae7a8b
--- /dev/null
+++ b/result/per_paper/2311.11413/accuracy_efficiency.csv
@@ -0,0 +1,39 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,BasicMotions,LPTM,accuracy,1.00,
+accuracy,FaceDetection,LPTM,accuracy,0.79,
+accuracy,FingerMovements,LPTM,accuracy,0.78,
+accuracy,PEMS-SF,LPTM,accuracy,0.93,
+accuracy,RacketSports,LPTM,accuracy,0.93,
+accuracy,EigenWorms,LPTM,accuracy,0.94,
+accuracy,ArticularyWordRecognition,LPTM,accuracy,0.98,
+accuracy,AtrialFibrillation,LPTM,accuracy,0.93,
+accuracy,CharacterTrajectories,LPTM,accuracy,0.98,
+accuracy,Cricket,LPTM,accuracy,0.99,
+accuracy,DuckGeese,LPTM,accuracy,0.79,
+accuracy,Epilepsy,LPTM,accuracy,0.97,
+accuracy,ERing,LPTM,accuracy,0.97,
+accuracy,EthanolConcentration,LPTM,accuracy,0.53,
+accuracy,HandMovementDirection,LPTM,accuracy,0.53,
+accuracy,Handwriting,LPTM,accuracy,0.51,
+accuracy,Heartbeat,LPTM,accuracy,0.74,
+accuracy,InsectWingbeat,LPTM,accuracy,0.72,
+accuracy,JapaneseVowels,LPTM,accuracy,0.98,
+accuracy,Libras,LPTM,accuracy,0.95,
+accuracy,LSST,LPTM,accuracy,0.98,
+accuracy,MotorImagery,LPTM,accuracy,0.57,
+accuracy,NATOPS,LPTM,accuracy,0.94,
+accuracy,PenDigits,LPTM,accuracy,0.92,
+accuracy,Phoneme,LPTM,accuracy,0.32,
+accuracy,SelfRegulation,LPTM,accuracy,0.92,
+accuracy,SpokenArabicDigits,LPTM,accuracy,1.00,
+accuracy,StandWalkJump,LPTM,accuracy,0.58,
+accuracy,UWaveGesture,LPTM,accuracy,0.94,
+accuracy,PAMAP2,LPTM,accuracy,0.97,
+accuracy,OpportunityGestures,LPTM,accuracy,0.92,
+accuracy,OpportunityLocomotion,LPTM,accuracy,0.89,
+accuracy,SelfRegulationSCP2,LPTM,accuracy,0.691,
+accuracy,Occupancy,LPTM,accuracy,0.836,
+accuracy,MosquitoSound,LPTM,accuracy,0.715,
+accuracy,TS2Vec,LPTM,RMSE,2,
+accuracy,TS2Vec,LPTM,RMSE,2,
+accuracy,TS2Vec,LPTM,RMSE,2,
diff --git a/result/per_paper/2311.11413/accuracy_efficiency_traced.csv b/result/per_paper/2311.11413/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..351d93d7ffba87b0c20b71760f8209ef385ea649
--- /dev/null
+++ b/result/per_paper/2311.11413/accuracy_efficiency_traced.csv
@@ -0,0 +1,39 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,BasicMotions,LPTM,accuracy,1.00,,0,1,10
+accuracy,FaceDetection,LPTM,accuracy,0.79,,0,2,10
+accuracy,FingerMovements,LPTM,accuracy,0.78,,0,3,10
+accuracy,PEMS-SF,LPTM,accuracy,0.93,,0,4,10
+accuracy,RacketSports,LPTM,accuracy,0.93,,0,5,10
+accuracy,EigenWorms,LPTM,accuracy,0.94,,0,6,10
+accuracy,ArticularyWordRecognition,LPTM,accuracy,0.98,,0,7,10
+accuracy,AtrialFibrillation,LPTM,accuracy,0.93,,0,8,10
+accuracy,CharacterTrajectories,LPTM,accuracy,0.98,,0,9,10
+accuracy,Cricket,LPTM,accuracy,0.99,,0,10,10
+accuracy,DuckGeese,LPTM,accuracy,0.79,,0,11,10
+accuracy,Epilepsy,LPTM,accuracy,0.97,,0,12,10
+accuracy,ERing,LPTM,accuracy,0.97,,0,13,10
+accuracy,EthanolConcentration,LPTM,accuracy,0.53,,0,14,10
+accuracy,HandMovementDirection,LPTM,accuracy,0.53,,0,15,10
+accuracy,Handwriting,LPTM,accuracy,0.51,,0,16,10
+accuracy,Heartbeat,LPTM,accuracy,0.74,,0,17,10
+accuracy,InsectWingbeat,LPTM,accuracy,0.72,,0,18,10
+accuracy,JapaneseVowels,LPTM,accuracy,0.98,,0,19,10
+accuracy,Libras,LPTM,accuracy,0.95,,0,20,10
+accuracy,LSST,LPTM,accuracy,0.98,,0,21,10
+accuracy,MotorImagery,LPTM,accuracy,0.57,,0,22,10
+accuracy,NATOPS,LPTM,accuracy,0.94,,0,23,10
+accuracy,PenDigits,LPTM,accuracy,0.92,,0,24,10
+accuracy,Phoneme,LPTM,accuracy,0.32,,0,25,10
+accuracy,SelfRegulation,LPTM,accuracy,0.92,,0,26,10
+accuracy,SpokenArabicDigits,LPTM,accuracy,1.00,,0,27,10
+accuracy,StandWalkJump,LPTM,accuracy,0.58,,0,28,10
+accuracy,UWaveGesture,LPTM,accuracy,0.94,,0,29,10
+accuracy,PAMAP2,LPTM,accuracy,0.97,,0,30,10
+accuracy,OpportunityGestures,LPTM,accuracy,0.92,,0,31,10
+accuracy,OpportunityLocomotion,LPTM,accuracy,0.89,,0,32,10
+accuracy,SelfRegulationSCP2,LPTM,accuracy,0.691,,0,33,10
+accuracy,Occupancy,LPTM,accuracy,0.836,,0,34,10
+accuracy,MosquitoSound,LPTM,accuracy,0.715,,0,35,10
+accuracy,TS2Vec,LPTM,RMSE,2,,1,19,0
+accuracy,TS2Vec,LPTM,RMSE,2,,2,15,0
+accuracy,TS2Vec,LPTM,RMSE,2,,4,6,0
diff --git a/result/per_paper/2311.11413/components_architecture.csv b/result/per_paper/2311.11413/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5fbfc0cbc2e6e02e54233719c9f5ebea583bc53e
--- /dev/null
+++ b/result/per_paper/2311.11413/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Transformer-based Architecture,"The core model structure based on the transformer architecture, adapted for time-series data.",reused_cited,"Vaswani et al., 2017",LPTM uses a simple transformer-based architecture.
+Adaptive Segmentation Module,"A novel module that dynamically identifies optimal segmentation strategies for time-series data during pre-training, tailored to domain-specific characteristics.",proposed_here,,We propose a novel adaptive segmentation module that segments the time-series of each domain based on how well it performs on self-supervised pre-training.
+Segmentation Scoring Mechanism,A method within the adaptive segmentation module that evaluates segmentation effectiveness using losses from self-supervised learning tasks.,proposed_here,,The segmentation module uses a novel scoring mechanism during pre-training to identify an effective segmentation strategy for a domain.
+Masking-based Self-Supervised Pre-training,"A pre-training strategy where patches of input data are masked, and the model is trained to reconstruct them, enabling learning of temporal patterns without labeled data.",reused_cited,"Devlin et al., 2019",These tasks mask patches of the input data and train the model to...
+Time-Series Tokenization via Segmentation,"A method of converting raw time-series data into tokens by segmenting the series into meaningful chunks, replacing per-time-step inputs.",proposed_here,,We propose a novel method of adaptive segmentation that automatically identifies optimal dataset-specific segmentation strategy during pre-training.
+Multi-Domain Pre-Training Dataset,"A collection of heterogeneous time-series datasets from diverse domains (e.g., epidemiology, energy, economics) used to train the model for cross-domain generalization.",proposed_here,,"The entire set of heterogeneous multi-domain pre-train dataset is denoted as $D_{pre} = \{D'_{1}, D'_{2}, \ldots, D'_{K}\}$."
diff --git a/result/per_paper/2311.11413/computational.csv b/result/per_paper/2311.11413/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f8ad2f4239b76c843d707698033591a1d3e1d592
--- /dev/null
+++ b/result/per_paper/2311.11413/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware specifications (e.g., GPU/TPU types) are mentioned for LPTM."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not reported."
+training_cost,,,not_reported,"Training cost (e.g., computational resources, energy, or monetary cost) is not mentioned."
+training_batch_size,,,not_reported,Batch size during training is not specified.
+training_steps_or_epochs,,,not_reported,Number of training steps or epochs is not reported.
+precision,,,not_reported,"Training or inference precision (e.g., FP16, FP32) is not mentioned."
+inference_latency,,,not_reported,Latency per inference is not provided.
+inference_throughput,,,not_reported,"Throughput (e.g., samples/second) is not reported."
+peak_memory,,,not_reported,Peak memory usage during training or inference is not mentioned.
+flops_or_macs,,,not_reported,FLOPs or MACs (computational complexity) is not provided.
+num_inference_samples,,,not_reported,Number of samples used for inference evaluation is not specified.
+params,,,not_reported,Model parameter count is not reported.
+context_lengths_evaluated,,,not_reported,"Context lengths (e.g., sequence lengths) evaluated during testing are not mentioned."
+horizon_lengths_evaluated,,,not_reported,Forecasting horizons evaluated are not specified.
+inference_batch_size,,,not_reported,Batch size during inference is not reported.
diff --git a/result/per_paper/2401.03955/accuracy_efficiency.csv b/result/per_paper/2401.03955/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d0094eaa186665e7cbeefa81ea6a711f96024a85
--- /dev/null
+++ b/result/per_paper/2401.03955/accuracy_efficiency.csv
@@ -0,0 +1,220 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTH1,TTM_B,MSE,0.36,
+accuracy,ETTH1,TTM_E,MSE,0.362,
+accuracy,ETTH1,TTM_A,MSE,0.363,
+accuracy,ETTH2,TTM_B,MSE,0.269,
+accuracy,ETTH2,TTM_E,MSE,0.273,
+accuracy,ETTH2,TTM_A,MSE,0.262,
+accuracy,ETTM1,TTM_B,MSE,0.291,
+accuracy,ETTM1,TTM_E,MSE,0.293,
+accuracy,ETTM1,TTM_A,MSE,0.283,
+accuracy,ETTM2,TTM_B,MSE,0.164,
+accuracy,ETTM2,TTM_E,MSE,0.158,
+accuracy,ETTM2,TTM_A,MSE,0.158,
+accuracy,Weather,TTM_B,MSE,0.146,
+accuracy,Weather,TTM_E,MSE,0.154,
+accuracy,Weather,TTM_A,MSE,0.149,
+accuracy,Electricity,TTM_B,MSE,0.129,
+accuracy,Electricity,TTM_E,MSE,0.129,
+accuracy,Electricity,TTM_A,MSE,0.128,
+accuracy,Traffic,TTM_B,MSE,0.368,
+accuracy,Traffic,TTM_E,MSE,0.372,
+accuracy,Traffic,TTM_A,MSE,0.352,
+accuracy,ETTH1,TTM_B,MSE,0.373,
+accuracy,ETTH1,TTM_E,MSE,0.36,
+accuracy,ETTH1,TTM_A,MSE,0.362,
+accuracy,ETTH2,TTM_B,MSE,0.283,
+accuracy,ETTH2,TTM_E,MSE,0.269,
+accuracy,ETTH2,TTM_A,MSE,0.273,
+accuracy,ETTM1,TTM_B,MSE,0.286,
+accuracy,ETTM1,TTM_E,MSE,0.291,
+accuracy,ETTM1,TTM_A,MSE,0.293,
+accuracy,ETTM2,TTM_B,MSE,0.165,
+accuracy,ETTM2,TTM_E,MSE,0.164,
+accuracy,ETTM2,TTM_A,MSE,0.158,
+accuracy,Weather,TTM_B,MSE,0.144,
+accuracy,Weather,TTM_E,MSE,0.146,
+accuracy,Weather,TTM_A,MSE,0.154,
+accuracy,Electricity,TTM_B,MSE,0.13,
+accuracy,Electricity,TTM_E,MSE,0.129,
+accuracy,Electricity,TTM_A,MSE,0.129,
+accuracy,Traffic,TTM_B,MSE,0.367,
+accuracy,Traffic,TTM_E,MSE,0.368,
+accuracy,Traffic,TTM_A,MSE,0.372,
+accuracy,ETTH1,TTM_B,MAE,0.195,
+accuracy,ETTH1,TTM_E,MAE,0.243,
+accuracy,ETTH1,TTM_A,MAE,0.217,
+accuracy,ETTH2,TTM_B,MAE,0.12,
+accuracy,ETTH2,TTM_E,MAE,0.149,
+accuracy,ETTH2,TTM_A,MAE,0.158,
+accuracy,ETTM1,TTM_B,MAE,0.141,
+accuracy,ETTM1,TTM_E,MAE,0.247,
+accuracy,ETTM1,TTM_A,MAE,0.187,
+accuracy,ETTM2,TTM_B,MAE,0.04,
+accuracy,ETTM2,TTM_E,MAE,0.063,
+accuracy,ETTM2,TTM_A,MAE,0.068,
+accuracy,Weather,TTM_B,MAE,0.015,
+accuracy,Weather,TTM_E,MAE,0.015,
+accuracy,Weather,TTM_A,MAE,0.016,
+accuracy,Electricity,TTM_B,MAE,0.409,
+accuracy,Electricity,TTM_E,MAE,0.421,
+accuracy,Electricity,TTM_A,MAE,0.433,
+accuracy,Traffic,TTM_B,MAE,0.231,
+accuracy,Traffic,TTM_E,MAE,0.232,
+accuracy,Traffic,TTM_A,MAE,0.239,
+accuracy,ETTH1,TTM-B,MAE,0.364,
+accuracy,ETTH1,TTM-E,MAE,0.363,
+accuracy,ETTH1,TTM-A,MAE,0.359,
+accuracy,ETTH2,TTM-B,MAE,0.277,
+accuracy,ETTH2,TTM-E,MAE,0.271,
+accuracy,ETTH2,TTM-A,MAE,0.267,
+accuracy,ETTM1,TTM-B,MAE,0.313,
+accuracy,ETTM1,TTM-E,MAE,0.326,
+accuracy,ETTM1,TTM-A,MAE,0.317,
+accuracy,ETTM2,TTM-B,MAE,0.171,
+accuracy,ETTM2,TTM-E,MAE,0.178,
+accuracy,ETTM2,TTM-A,MAE,0.17,
+accuracy,Weather,TTM-B,MAE,0.154,
+accuracy,Weather,TTM-E,MAE,0.162,
+accuracy,Weather,TTM-A,MAE,0.155,
+accuracy,Electricity,TTM-B,MAE,0.146,
+accuracy,Electricity,TTM-E,MAE,0.15,
+accuracy,Electricity,TTM-A,MAE,0.141,
+accuracy,Traffic,TTM-B,MAE,0.411,
+accuracy,Traffic,TTM-E,MAE,0.411,
+accuracy,Traffic,TTM-A,MAE,0.469,
+accuracy,ETTH1,TTM-B,MAE,0.364,
+accuracy,ETTH1,TTM-E,MAE,0.363,
+accuracy,ETTH1,TTM-A,MAE,0.359,
+accuracy,ETTH2,TTM-B,MAE,0.277,
+accuracy,ETTH2,TTM-E,MAE,0.271,
+accuracy,ETTH2,TTM-A,MAE,0.264,
+accuracy,ETTM1,TTM-B,MAE,0.322,
+accuracy,ETTM1,TTM-E,MAE,0.327,
+accuracy,ETTM1,TTM-A,MAE,0.318,
+accuracy,ETTM2,TTM-B,MAE,0.171,
+accuracy,ETTM2,TTM-E,MAE,0.178,
+accuracy,ETTM2,TTM-A,MAE,0.169,
+accuracy,Weather,TTM-B,MAE,0.158,
+accuracy,Weather,TTM-E,MAE,0.166,
+accuracy,Weather,TTM-A,MAE,0.159,
+accuracy,Electricity,TTM-B,MAE,0.166,
+accuracy,Electricity,TTM-E,MAE,0.157,
+accuracy,Electricity,TTM-A,MAE,0.152,
+accuracy,ETTH1,TTM_Q,FL,0.365,
+accuracy,ETTH1,TTM_B,FL,0.364,
+accuracy,ETTH1,TTM_E,FL,0.363,
+accuracy,ETTH1,TTM_A,FL,0.359,
+accuracy,ETTH2,TTM_Q,FL,0.285,
+accuracy,ETTH2,TTM_B,FL,0.277,
+accuracy,ETTH2,TTM_E,FL,0.271,
+accuracy,ETTH2,TTM_A,FL,0.264,
+accuracy,ETTM1,TTM_Q,FL,0.413,
+accuracy,ETTM1,TTM_B,FL,0.322,
+accuracy,ETTM1,TTM_E,FL,0.327,
+accuracy,ETTM1,TTM_A,FL,0.318,
+accuracy,ETTM2,TTM_Q,FL,0.187,
+accuracy,ETTM2,TTM_B,FL,0.171,
+accuracy,ETTM2,TTM_E,FL,0.178,
+accuracy,ETTM2,TTM_A,FL,0.169,
+accuracy,Weather,TTM_Q,FL,0.154,
+accuracy,Weather,TTM_B,FL,0.158,
+accuracy,Weather,TTM_E,FL,0.166,
+accuracy,Weather,TTM_A,FL,0.159,
+accuracy,Electricity,TTM_Q,FL,0.169,
+accuracy,Electricity,TTM_B,FL,0.166,
+accuracy,Electricity,TTM_E,FL,0.157,
+accuracy,Electricity,TTM_A,FL,0.152,
+accuracy,Traffic,TTM_Q,FL,0.518,
+accuracy,Traffic,TTM_B,FL,0.514,
+accuracy,Traffic,TTM_E,FL,0.476,
+accuracy,Traffic,TTM_A,FL,0.462,
+accuracy,Model Size,TTM_Q,FL,1,
+accuracy,Model Size,TTM_B,FL,4,
+accuracy,Model Size,TTM_E,FL,5,
+accuracy,ETTH1,TTM_Q,MAE,0.366,
+accuracy,ETTH1,TTM_B,MAE,0.364,
+accuracy,ETTH1,TTM_E,MAE,0.363,
+accuracy,ETTH1,TTM_A,MAE,0.359,
+accuracy,ETTH2,TTM_Q,MAE,0.282,
+accuracy,ETTH2,TTM_B,MAE,0.277,
+accuracy,ETTH2,TTM_E,MAE,0.271,
+accuracy,ETTH2,TTM_A,MAE,0.267,
+accuracy,ETTM1,TTM_Q,MAE,0.359,
+accuracy,ETTM1,TTM_B,MAE,0.313,
+accuracy,ETTM1,TTM_E,MAE,0.326,
+accuracy,ETTM1,TTM_A,MAE,0.317,
+accuracy,ETTM2,TTM_Q,MAE,0.174,
+accuracy,ETTM2,TTM_B,MAE,0.171,
+accuracy,ETTM2,TTM_E,MAE,0.178,
+accuracy,ETTM2,TTM_A,MAE,0.17,
+accuracy,Weather,TTM_Q,MAE,0.152,
+accuracy,Weather,TTM_B,MAE,0.154,
+accuracy,Weather,TTM_E,MAE,0.162,
+accuracy,Weather,TTM_A,MAE,0.155,
+accuracy,Electricity,TTM_Q,MAE,0.142,
+accuracy,Electricity,TTM_B,MAE,0.146,
+accuracy,Electricity,TTM_E,MAE,0.15,
+accuracy,Electricity,TTM_A,MAE,0.141,
+accuracy,Traffic,TTM_Q,MAE,0.401,
+accuracy,Traffic,TTM_B,MAE,0.411,
+accuracy,Traffic,TTM_E,MAE,0.411,
+accuracy,Traffic,TTM_A,MAE,0.469,
+accuracy,Model Size,TTM_Q,MAE,1,
+accuracy,Model Size,TTM_B,MAE,4,
+accuracy,Model Size,TTM_E,MAE,5,
+accuracy,ETTH1,TTM_Q,MAE,0.398,
+accuracy,ETTH1,TTM_B,MAE,0.397,
+accuracy,ETTH1,TTM_E,MAE,0.402,
+accuracy,ETTH1,TTM_A,MAE,0.402,
+accuracy,ETTH2,TTM_Q,MAE,0.347,
+accuracy,ETTH2,TTM_B,MAE,0.338,
+accuracy,ETTH2,TTM_E,MAE,0.339,
+accuracy,ETTH2,TTM_A,MAE,0.332,
+accuracy,ETTM1,TTM_Q,MAE,0.35,
+accuracy,ETTM1,TTM_B,MAE,0.35,
+accuracy,ETTM1,TTM_E,MAE,0.35,
+accuracy,ETTM1,TTM_A,MAE,0.34,
+accuracy,ETTM2,TTM_Q,MAE,0.253,
+accuracy,ETTM2,TTM_B,MAE,0.252,
+accuracy,ETTM2,TTM_E,MAE,0.245,
+accuracy,ETTM2,TTM_A,MAE,0.252,
+accuracy,Weather,TTM_Q,MAE,0.224,
+accuracy,Weather,TTM_B,MAE,0.225,
+accuracy,Weather,TTM_E,MAE,0.234,
+accuracy,Weather,TTM_A,MAE,0.225,
+accuracy,Electricity,TTM_Q,MAE,0.161,
+accuracy,Electricity,TTM_B,MAE,0.16,
+accuracy,Electricity,TTM_E,MAE,0.158,
+accuracy,Electricity,TTM_A,MAE,0.156,
+accuracy,Traffic,TTM_Q,MAE,0.4,
+accuracy,Traffic,TTM_B,MAE,0.399,
+accuracy,Traffic,TTM_E,MAE,0.385,
+accuracy,Traffic,TTM_A,MAE,0.376,
+accuracy,ETTH1,TTM-B,MAE,0.394,
+accuracy,ETTH1,TTM-E,MAE,0.404,
+accuracy,ETTH1,TTM-A,MAE,0.4,
+accuracy,ETTH2,TTM-B,MAE,0.345,
+accuracy,ETTH2,TTM-E,MAE,0.335,
+accuracy,ETTH2,TTM-A,MAE,0.333,
+accuracy,ETTM1,TTM-B,MAE,0.386,
+accuracy,ETTM1,TTM-E,MAE,0.38,
+accuracy,ETTM1,TTM-A,MAE,0.362,
+accuracy,ETTM2,TTM-B,MAE,0.281,
+accuracy,ETTM2,TTM-E,MAE,0.271,
+accuracy,ETTM2,TTM-A,MAE,0.252,
+accuracy,Weather,TTM-B,MAE,0.237,
+accuracy,Weather,TTM-E,MAE,0.238,
+accuracy,Weather,TTM-A,MAE,0.231,
+accuracy,Electricity,TTM-B,MAE,0.205,
+accuracy,Electricity,TTM-E,MAE,0.194,
+accuracy,Electricity,TTM-A,MAE,0.192,
+accuracy,$TTM_B$ f-imp(%),TTM-B,MAE,6,
+accuracy,$TTM_B$ f-imp(%),TTM-E,MAE,1,
+accuracy,$TTM_B$ f-imp(%),TTM-A,MAE,4,
+accuracy,$TTM_E$ f-imp(%),TTM-B,MAE,7,
+accuracy,$TTM_E$ f-imp(%),TTM-E,MAE,1,
+accuracy,$TTM_E$ f-imp(%),TTM-A,MAE,6,
+accuracy,$TTM_A$ f-imp(%),TTM-B,MAE,10,
+accuracy,$TTM_A$ f-imp(%),TTM-E,MAE,4,
+accuracy,$TTM_A$ f-imp(%),TTM-A,MAE,9,
diff --git a/result/per_paper/2401.03955/accuracy_efficiency_traced.csv b/result/per_paper/2401.03955/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..3424e2858f8d796d1050d535d0f825960d7993b3
--- /dev/null
+++ b/result/per_paper/2401.03955/accuracy_efficiency_traced.csv
@@ -0,0 +1,220 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTH1,TTM_B,MSE,0.36,,0,2,2
+accuracy,ETTH1,TTM_E,MSE,0.362,,0,2,3
+accuracy,ETTH1,TTM_A,MSE,0.363,,0,2,4
+accuracy,ETTH2,TTM_B,MSE,0.269,,0,4,2
+accuracy,ETTH2,TTM_E,MSE,0.273,,0,4,3
+accuracy,ETTH2,TTM_A,MSE,0.262,,0,4,4
+accuracy,ETTM1,TTM_B,MSE,0.291,,0,6,2
+accuracy,ETTM1,TTM_E,MSE,0.293,,0,6,3
+accuracy,ETTM1,TTM_A,MSE,0.283,,0,6,4
+accuracy,ETTM2,TTM_B,MSE,0.164,,0,8,2
+accuracy,ETTM2,TTM_E,MSE,0.158,,0,8,3
+accuracy,ETTM2,TTM_A,MSE,0.158,,0,8,4
+accuracy,Weather,TTM_B,MSE,0.146,,0,10,2
+accuracy,Weather,TTM_E,MSE,0.154,,0,10,3
+accuracy,Weather,TTM_A,MSE,0.149,,0,10,4
+accuracy,Electricity,TTM_B,MSE,0.129,,0,12,2
+accuracy,Electricity,TTM_E,MSE,0.129,,0,12,3
+accuracy,Electricity,TTM_A,MSE,0.128,,0,12,4
+accuracy,Traffic,TTM_B,MSE,0.368,,0,14,2
+accuracy,Traffic,TTM_E,MSE,0.372,,0,14,3
+accuracy,Traffic,TTM_A,MSE,0.352,,0,14,4
+accuracy,ETTH1,TTM_B,MSE,0.373,,0,18,2
+accuracy,ETTH1,TTM_E,MSE,0.36,,0,18,3
+accuracy,ETTH1,TTM_A,MSE,0.362,,0,18,4
+accuracy,ETTH2,TTM_B,MSE,0.283,,0,22,2
+accuracy,ETTH2,TTM_E,MSE,0.269,,0,22,3
+accuracy,ETTH2,TTM_A,MSE,0.273,,0,22,4
+accuracy,ETTM1,TTM_B,MSE,0.286,,0,26,2
+accuracy,ETTM1,TTM_E,MSE,0.291,,0,26,3
+accuracy,ETTM1,TTM_A,MSE,0.293,,0,26,4
+accuracy,ETTM2,TTM_B,MSE,0.165,,0,30,2
+accuracy,ETTM2,TTM_E,MSE,0.164,,0,30,3
+accuracy,ETTM2,TTM_A,MSE,0.158,,0,30,4
+accuracy,Weather,TTM_B,MSE,0.144,,0,34,2
+accuracy,Weather,TTM_E,MSE,0.146,,0,34,3
+accuracy,Weather,TTM_A,MSE,0.154,,0,34,4
+accuracy,Electricity,TTM_B,MSE,0.13,,0,38,2
+accuracy,Electricity,TTM_E,MSE,0.129,,0,38,3
+accuracy,Electricity,TTM_A,MSE,0.129,,0,38,4
+accuracy,Traffic,TTM_B,MSE,0.367,,0,42,2
+accuracy,Traffic,TTM_E,MSE,0.368,,0,42,3
+accuracy,Traffic,TTM_A,MSE,0.372,,0,42,4
+accuracy,ETTH1,TTM_B,MAE,0.195,,1,1,2
+accuracy,ETTH1,TTM_E,MAE,0.243,,1,1,3
+accuracy,ETTH1,TTM_A,MAE,0.217,,1,1,4
+accuracy,ETTH2,TTM_B,MAE,0.12,,1,6,2
+accuracy,ETTH2,TTM_E,MAE,0.149,,1,6,3
+accuracy,ETTH2,TTM_A,MAE,0.158,,1,6,4
+accuracy,ETTM1,TTM_B,MAE,0.141,,1,11,2
+accuracy,ETTM1,TTM_E,MAE,0.247,,1,11,3
+accuracy,ETTM1,TTM_A,MAE,0.187,,1,11,4
+accuracy,ETTM2,TTM_B,MAE,0.04,,1,16,2
+accuracy,ETTM2,TTM_E,MAE,0.063,,1,16,3
+accuracy,ETTM2,TTM_A,MAE,0.068,,1,16,4
+accuracy,Weather,TTM_B,MAE,0.015,,1,21,2
+accuracy,Weather,TTM_E,MAE,0.015,,1,21,3
+accuracy,Weather,TTM_A,MAE,0.016,,1,21,4
+accuracy,Electricity,TTM_B,MAE,0.409,,1,26,2
+accuracy,Electricity,TTM_E,MAE,0.421,,1,26,3
+accuracy,Electricity,TTM_A,MAE,0.433,,1,26,4
+accuracy,Traffic,TTM_B,MAE,0.231,,1,31,2
+accuracy,Traffic,TTM_E,MAE,0.232,,1,31,3
+accuracy,Traffic,TTM_A,MAE,0.239,,1,31,4
+accuracy,ETTH1,TTM-B,MAE,0.364,,2,2,2
+accuracy,ETTH1,TTM-E,MAE,0.363,,2,2,3
+accuracy,ETTH1,TTM-A,MAE,0.359,,2,2,4
+accuracy,ETTH2,TTM-B,MAE,0.277,,2,6,2
+accuracy,ETTH2,TTM-E,MAE,0.271,,2,6,3
+accuracy,ETTH2,TTM-A,MAE,0.267,,2,6,4
+accuracy,ETTM1,TTM-B,MAE,0.313,,2,10,2
+accuracy,ETTM1,TTM-E,MAE,0.326,,2,10,3
+accuracy,ETTM1,TTM-A,MAE,0.317,,2,10,4
+accuracy,ETTM2,TTM-B,MAE,0.171,,2,14,2
+accuracy,ETTM2,TTM-E,MAE,0.178,,2,14,3
+accuracy,ETTM2,TTM-A,MAE,0.17,,2,14,4
+accuracy,Weather,TTM-B,MAE,0.154,,2,18,2
+accuracy,Weather,TTM-E,MAE,0.162,,2,18,3
+accuracy,Weather,TTM-A,MAE,0.155,,2,18,4
+accuracy,Electricity,TTM-B,MAE,0.146,,2,22,2
+accuracy,Electricity,TTM-E,MAE,0.15,,2,22,3
+accuracy,Electricity,TTM-A,MAE,0.141,,2,22,4
+accuracy,Traffic,TTM-B,MAE,0.411,,2,26,2
+accuracy,Traffic,TTM-E,MAE,0.411,,2,26,3
+accuracy,Traffic,TTM-A,MAE,0.469,,2,26,4
+accuracy,ETTH1,TTM-B,MAE,0.364,,3,1,2
+accuracy,ETTH1,TTM-E,MAE,0.363,,3,1,3
+accuracy,ETTH1,TTM-A,MAE,0.359,,3,1,4
+accuracy,ETTH2,TTM-B,MAE,0.277,,3,5,2
+accuracy,ETTH2,TTM-E,MAE,0.271,,3,5,3
+accuracy,ETTH2,TTM-A,MAE,0.264,,3,5,4
+accuracy,ETTM1,TTM-B,MAE,0.322,,3,9,2
+accuracy,ETTM1,TTM-E,MAE,0.327,,3,9,3
+accuracy,ETTM1,TTM-A,MAE,0.318,,3,9,4
+accuracy,ETTM2,TTM-B,MAE,0.171,,3,13,2
+accuracy,ETTM2,TTM-E,MAE,0.178,,3,13,3
+accuracy,ETTM2,TTM-A,MAE,0.169,,3,13,4
+accuracy,Weather,TTM-B,MAE,0.158,,3,17,2
+accuracy,Weather,TTM-E,MAE,0.166,,3,17,3
+accuracy,Weather,TTM-A,MAE,0.159,,3,17,4
+accuracy,Electricity,TTM-B,MAE,0.166,,3,21,2
+accuracy,Electricity,TTM-E,MAE,0.157,,3,21,3
+accuracy,Electricity,TTM-A,MAE,0.152,,3,21,4
+accuracy,ETTH1,TTM_Q,FL,0.365,,4,1,2
+accuracy,ETTH1,TTM_B,FL,0.364,,4,1,3
+accuracy,ETTH1,TTM_E,FL,0.363,,4,1,4
+accuracy,ETTH1,TTM_A,FL,0.359,,4,1,5
+accuracy,ETTH2,TTM_Q,FL,0.285,,4,5,2
+accuracy,ETTH2,TTM_B,FL,0.277,,4,5,3
+accuracy,ETTH2,TTM_E,FL,0.271,,4,5,4
+accuracy,ETTH2,TTM_A,FL,0.264,,4,5,5
+accuracy,ETTM1,TTM_Q,FL,0.413,,4,9,2
+accuracy,ETTM1,TTM_B,FL,0.322,,4,9,3
+accuracy,ETTM1,TTM_E,FL,0.327,,4,9,4
+accuracy,ETTM1,TTM_A,FL,0.318,,4,9,5
+accuracy,ETTM2,TTM_Q,FL,0.187,,4,13,2
+accuracy,ETTM2,TTM_B,FL,0.171,,4,13,3
+accuracy,ETTM2,TTM_E,FL,0.178,,4,13,4
+accuracy,ETTM2,TTM_A,FL,0.169,,4,13,5
+accuracy,Weather,TTM_Q,FL,0.154,,4,17,2
+accuracy,Weather,TTM_B,FL,0.158,,4,17,3
+accuracy,Weather,TTM_E,FL,0.166,,4,17,4
+accuracy,Weather,TTM_A,FL,0.159,,4,17,5
+accuracy,Electricity,TTM_Q,FL,0.169,,4,21,2
+accuracy,Electricity,TTM_B,FL,0.166,,4,21,3
+accuracy,Electricity,TTM_E,FL,0.157,,4,21,4
+accuracy,Electricity,TTM_A,FL,0.152,,4,21,5
+accuracy,Traffic,TTM_Q,FL,0.518,,4,25,2
+accuracy,Traffic,TTM_B,FL,0.514,,4,25,3
+accuracy,Traffic,TTM_E,FL,0.476,,4,25,4
+accuracy,Traffic,TTM_A,FL,0.462,,4,25,5
+accuracy,Model Size,TTM_Q,FL,1,,4,29,2
+accuracy,Model Size,TTM_B,FL,4,,4,29,3
+accuracy,Model Size,TTM_E,FL,5,,4,29,4
+accuracy,ETTH1,TTM_Q,MAE,0.366,,5,1,2
+accuracy,ETTH1,TTM_B,MAE,0.364,,5,1,3
+accuracy,ETTH1,TTM_E,MAE,0.363,,5,1,4
+accuracy,ETTH1,TTM_A,MAE,0.359,,5,1,5
+accuracy,ETTH2,TTM_Q,MAE,0.282,,5,5,2
+accuracy,ETTH2,TTM_B,MAE,0.277,,5,5,3
+accuracy,ETTH2,TTM_E,MAE,0.271,,5,5,4
+accuracy,ETTH2,TTM_A,MAE,0.267,,5,5,5
+accuracy,ETTM1,TTM_Q,MAE,0.359,,5,9,2
+accuracy,ETTM1,TTM_B,MAE,0.313,,5,9,3
+accuracy,ETTM1,TTM_E,MAE,0.326,,5,9,4
+accuracy,ETTM1,TTM_A,MAE,0.317,,5,9,5
+accuracy,ETTM2,TTM_Q,MAE,0.174,,5,13,2
+accuracy,ETTM2,TTM_B,MAE,0.171,,5,13,3
+accuracy,ETTM2,TTM_E,MAE,0.178,,5,13,4
+accuracy,ETTM2,TTM_A,MAE,0.17,,5,13,5
+accuracy,Weather,TTM_Q,MAE,0.152,,5,17,2
+accuracy,Weather,TTM_B,MAE,0.154,,5,17,3
+accuracy,Weather,TTM_E,MAE,0.162,,5,17,4
+accuracy,Weather,TTM_A,MAE,0.155,,5,17,5
+accuracy,Electricity,TTM_Q,MAE,0.142,,5,21,2
+accuracy,Electricity,TTM_B,MAE,0.146,,5,21,3
+accuracy,Electricity,TTM_E,MAE,0.15,,5,21,4
+accuracy,Electricity,TTM_A,MAE,0.141,,5,21,5
+accuracy,Traffic,TTM_Q,MAE,0.401,,5,25,2
+accuracy,Traffic,TTM_B,MAE,0.411,,5,25,3
+accuracy,Traffic,TTM_E,MAE,0.411,,5,25,4
+accuracy,Traffic,TTM_A,MAE,0.469,,5,25,5
+accuracy,Model Size,TTM_Q,MAE,1,,5,29,2
+accuracy,Model Size,TTM_B,MAE,4,,5,29,3
+accuracy,Model Size,TTM_E,MAE,5,,5,29,4
+accuracy,ETTH1,TTM_Q,MAE,0.398,,6,2,1
+accuracy,ETTH1,TTM_B,MAE,0.397,,6,2,2
+accuracy,ETTH1,TTM_E,MAE,0.402,,6,2,3
+accuracy,ETTH1,TTM_A,MAE,0.402,,6,2,4
+accuracy,ETTH2,TTM_Q,MAE,0.347,,6,3,1
+accuracy,ETTH2,TTM_B,MAE,0.338,,6,3,2
+accuracy,ETTH2,TTM_E,MAE,0.339,,6,3,3
+accuracy,ETTH2,TTM_A,MAE,0.332,,6,3,4
+accuracy,ETTM1,TTM_Q,MAE,0.35,,6,4,1
+accuracy,ETTM1,TTM_B,MAE,0.35,,6,4,2
+accuracy,ETTM1,TTM_E,MAE,0.35,,6,4,3
+accuracy,ETTM1,TTM_A,MAE,0.34,,6,4,4
+accuracy,ETTM2,TTM_Q,MAE,0.253,,6,5,1
+accuracy,ETTM2,TTM_B,MAE,0.252,,6,5,2
+accuracy,ETTM2,TTM_E,MAE,0.245,,6,5,3
+accuracy,ETTM2,TTM_A,MAE,0.252,,6,5,4
+accuracy,Weather,TTM_Q,MAE,0.224,,6,6,1
+accuracy,Weather,TTM_B,MAE,0.225,,6,6,2
+accuracy,Weather,TTM_E,MAE,0.234,,6,6,3
+accuracy,Weather,TTM_A,MAE,0.225,,6,6,4
+accuracy,Electricity,TTM_Q,MAE,0.161,,6,7,1
+accuracy,Electricity,TTM_B,MAE,0.16,,6,7,2
+accuracy,Electricity,TTM_E,MAE,0.158,,6,7,3
+accuracy,Electricity,TTM_A,MAE,0.156,,6,7,4
+accuracy,Traffic,TTM_Q,MAE,0.4,,6,8,1
+accuracy,Traffic,TTM_B,MAE,0.399,,6,8,2
+accuracy,Traffic,TTM_E,MAE,0.385,,6,8,3
+accuracy,Traffic,TTM_A,MAE,0.376,,6,8,4
+accuracy,ETTH1,TTM-B,MAE,0.394,,7,2,1
+accuracy,ETTH1,TTM-E,MAE,0.404,,7,2,2
+accuracy,ETTH1,TTM-A,MAE,0.4,,7,2,3
+accuracy,ETTH2,TTM-B,MAE,0.345,,7,3,1
+accuracy,ETTH2,TTM-E,MAE,0.335,,7,3,2
+accuracy,ETTH2,TTM-A,MAE,0.333,,7,3,3
+accuracy,ETTM1,TTM-B,MAE,0.386,,7,4,1
+accuracy,ETTM1,TTM-E,MAE,0.38,,7,4,2
+accuracy,ETTM1,TTM-A,MAE,0.362,,7,4,3
+accuracy,ETTM2,TTM-B,MAE,0.281,,7,5,1
+accuracy,ETTM2,TTM-E,MAE,0.271,,7,5,2
+accuracy,ETTM2,TTM-A,MAE,0.252,,7,5,3
+accuracy,Weather,TTM-B,MAE,0.237,,7,6,1
+accuracy,Weather,TTM-E,MAE,0.238,,7,6,2
+accuracy,Weather,TTM-A,MAE,0.231,,7,6,3
+accuracy,Electricity,TTM-B,MAE,0.205,,7,7,1
+accuracy,Electricity,TTM-E,MAE,0.194,,7,7,2
+accuracy,Electricity,TTM-A,MAE,0.192,,7,7,3
+accuracy,$TTM_B$ f-imp(%),TTM-B,MAE,6,,7,8,1
+accuracy,$TTM_B$ f-imp(%),TTM-E,MAE,1,,7,8,2
+accuracy,$TTM_B$ f-imp(%),TTM-A,MAE,4,,7,8,3
+accuracy,$TTM_E$ f-imp(%),TTM-B,MAE,7,,7,9,1
+accuracy,$TTM_E$ f-imp(%),TTM-E,MAE,1,,7,9,2
+accuracy,$TTM_E$ f-imp(%),TTM-A,MAE,6,,7,9,3
+accuracy,$TTM_A$ f-imp(%),TTM-B,MAE,10,,7,10,1
+accuracy,$TTM_A$ f-imp(%),TTM-E,MAE,4,,7,10,2
+accuracy,$TTM_A$ f-imp(%),TTM-A,MAE,9,,7,10,3
diff --git a/result/per_paper/2401.03955/components_architecture.csv b/result/per_paper/2401.03955/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b5195b2809231c363b9dba411b75822d0782c8df
--- /dev/null
+++ b/result/per_paper/2401.03955/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+MLPMixer Blocks,"A lightweight architectural component using multi-layer perceptron (MLP) operations to process time series data, replacing self-attention mechanisms in Transformers.",reused_cited,"TSMixer, 2021",TTM is based on the light-weight TSMixer architecture that uses MLPMixer blocks interleaved with simple gated attention as alternatives to the quadratic time-consuming self-attention blocks in Transformers.
+Gated Attention,A simplified attention mechanism that combines gating functions with attention to reduce computational complexity while maintaining performance.,reused_cited,"TSMixer, 2021",TTM is based on the light-weight TSMixer architecture that uses MLPMixer blocks interleaved with simple gated attention as alternatives to the quadratic time-consuming self-attention blocks in Transformers.
+Adaptive Patching,"A technique to dynamically partition time series data into variable-length patches during pre-training, enabling the model to handle diverse temporal resolutions.",proposed_here,,"TTM incorporates innovations like adaptive patching, diverse resolution sampling, and resolution prefix tuning to handle pre-training on varied dataset resolutions with minimal model capacity."
+Diverse Resolution Sampling,"A method to train the model on time series data with varying temporal resolutions (e.g., seconds to days) to improve generalization across different domains.",proposed_here,,"TTM incorporates innovations like adaptive patching, diverse resolution sampling, and resolution prefix tuning to handle pre-training on varied dataset resolutions with minimal model capacity."
+Resolution Prefix Tuning,A parameter-efficient technique to adapt the model to specific temporal resolutions during fine-tuning by modifying prefix embeddings.,proposed_here,,"TTM incorporates innovations like adaptive patching, diverse resolution sampling, and resolution prefix tuning to handle pre-training on varied dataset resolutions with minimal model capacity."
+Multi-level Modeling,A framework to capture cross-channel correlations (between target variables) and integrate exogenous signals (external influencing variables) during fine-tuning.,proposed_here,,"Additionally, it employs multi-level modeling to capture channel correlations and infuse exogenous signals during fine-tuning."
diff --git a/result/per_paper/2401.03955/computational.csv b/result/per_paper/2401.03955/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..baa740ed79017f8d8a607860b26f24b56eff31b1
--- /dev/null
+++ b/result/per_paper/2401.03955/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,CPU or GPU,not_applicable,stated,Outline of TTM's key capabilities: ... zero-shot inference and fine-tuning of TTM models can be easily executed with just one GPU or in CPU-only environments.
+num_devices,1,not_applicable,stated,Outline of TTM's key capabilities: ... zero-shot inference and fine-tuning of TTM models can be easily executed with just one GPU or in CPU-only environments.
+training_cost,,not_applicable,not_reported,Not mentioned in the text.
+training_batch_size,,not_applicable,not_reported,Not mentioned in the text.
+training_steps_or_epochs,,not_applicable,not_reported,Not mentioned in the text.
+precision,,not_applicable,not_reported,Not mentioned in the text.
+inference_latency,0.01,s,stated,Table: TTM_B | ... | CPU TIME (s): 0.01
+inference_throughput,,not_applicable,not_reported,Not mentioned in the text.
+peak_memory,0.06,GB,stated,Table: TTM_B | ... | MEM (GB): 0.06
+flops_or_macs,,not_applicable,not_reported,Not mentioned in the text.
+num_inference_samples,,not_applicable,not_reported,Not mentioned in the text.
+params,1,M,stated,Table: Size | 1M | 4M | 5M | ... for TTM_B
+context_lengths_evaluated,,not_applicable,not_reported,Not mentioned in the text.
+horizon_lengths_evaluated,,not_applicable,not_reported,Not mentioned in the text.
+inference_batch_size,,not_applicable,not_reported,Not mentioned in the text.
diff --git a/result/per_paper/2402.02368/accuracy_efficiency.csv b/result/per_paper/2402.02368/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..606b3c8118096296878e67d012fb19802299012a
--- /dev/null
+++ b/result/per_paper/2402.02368/accuracy_efficiency.csv
@@ -0,0 +1,108 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,
+accuracy,SOTA,Timer-12G,MAE,0.129,
+accuracy,SOTA,Timer-12G,MAE,0.149,
+accuracy,SOTA,Timer-12G,MAE,0.115,
+efficiency,MASK RATIO,Timer-12G,Δ%,25.0,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,ETTH1,Timer-12G,Δ%,0.278,
+efficiency,ETTH1,Timer-12G,Δ%,0.287,
+efficiency,ETTH1,Timer-12G,Δ%,0.297,
+efficiency,ETTH1,Timer-12G,Δ%,0.314,
+efficiency,ETTH2,Timer-12G,Δ%,0.166,
+efficiency,ETTH2,Timer-12G,Δ%,0.178,
+efficiency,ETTH2,Timer-12G,Δ%,0.190,
+efficiency,ETTH2,Timer-12G,Δ%,0.208,
+efficiency,ETTM1,Timer-12G,Δ%,0.328,
+efficiency,ETTM1,Timer-12G,Δ%,0.326,
+efficiency,ETTM1,Timer-12G,Δ%,0.360,
+efficiency,ETTM1,Timer-12G,Δ%,0.407,
+efficiency,ETTM2,Timer-12G,Δ%,0.133,
+efficiency,ETTM2,Timer-12G,Δ%,0.123,
+efficiency,ETTM2,Timer-12G,Δ%,0.136,
+efficiency,ETTM2,Timer-12G,Δ%,0.143,
+efficiency,ECL,Timer-12G,Δ%,0.130,
+efficiency,ECL,Timer-12G,Δ%,0.138,
+efficiency,ECL,Timer-12G,Δ%,0.149,
+efficiency,ECL,Timer-12G,Δ%,0.162,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.420,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.446,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.477,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.511,
+efficiency,WEATHER,Timer-12G,Δ%,0.129,
+efficiency,WEATHER,Timer-12G,Δ%,0.147,
+efficiency,WEATHER,Timer-12G,Δ%,0.125,
+efficiency,WEATHER,Timer-12G,Δ%,0.153,
+efficiency,PEMS03,Timer-12G,Δ%,0.120,
+efficiency,PEMS03,Timer-12G,Δ%,0.150,
+efficiency,PEMS03,Timer-12G,Δ%,0.198,
+efficiency,PEMS03,Timer-12G,Δ%,0.273,
+efficiency,PEMS04,Timer-12G,Δ%,0.146,
+efficiency,PEMS04,Timer-12G,Δ%,0.184,
+efficiency,PEMS04,Timer-12G,Δ%,0.236,
+efficiency,PEMS04,Timer-12G,Δ%,0.320,
+efficiency,PEMS07,Timer-12G,Δ%,0.125,
+efficiency,PEMS07,Timer-12G,Δ%,0.162,
+efficiency,PEMS07,Timer-12G,Δ%,0.214,
+efficiency,PEMS07,Timer-12G,Δ%,0.290,
+efficiency,PEMS08,Timer-12G,Δ%,0.139,
+efficiency,PEMS08,Timer-12G,Δ%,0.174,
+efficiency,PEMS08,Timer-12G,Δ%,0.236,
+efficiency,PEMS08,Timer-12G,Δ%,0.324,
+efficiency,MASK RATIO,Timer-12G,Δ%,25.0,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,
+efficiency,ETTH1,Timer-12G,Δ%,0.273,
+efficiency,ETTH1,Timer-12G,Δ%,0.283,
+efficiency,ETTH1,Timer-12G,Δ%,0.294,
+efficiency,ETTH1,Timer-12G,Δ%,0.312,
+efficiency,ETTH2,Timer-12G,Δ%,0.177,
+efficiency,ETTH2,Timer-12G,Δ%,0.186,
+efficiency,ETTH2,Timer-12G,Δ%,0.195,
+efficiency,ETTH2,Timer-12G,Δ%,0.209,
+efficiency,ETTM1,Timer-12G,Δ%,0.352,
+efficiency,ETTM1,Timer-12G,Δ%,0.345,
+efficiency,ETTM1,Timer-12G,Δ%,0.371,
+efficiency,ETTM1,Timer-12G,Δ%,0.413,
+efficiency,ETTM2,Timer-12G,Δ%,0.161,
+efficiency,ETTM2,Timer-12G,Δ%,0.171,
+efficiency,ETTM2,Timer-12G,Δ%,0.176,
+efficiency,ETTM2,Timer-12G,Δ%,0.158,
+efficiency,ECL,Timer-12G,Δ%,0.122,
+efficiency,ECL,Timer-12G,Δ%,0.130,
+efficiency,ECL,Timer-12G,Δ%,0.139,
+efficiency,ECL,Timer-12G,Δ%,0.152,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.392,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.414,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.443,
+efficiency,TRAFFIC,Timer-12G,Δ%,0.477,
+efficiency,WEATHER,Timer-12G,Δ%,0.157,
+efficiency,WEATHER,Timer-12G,Δ%,0.146,
+efficiency,WEATHER,Timer-12G,Δ%,0.147,
+efficiency,WEATHER,Timer-12G,Δ%,0.158,
+efficiency,PEMS03,Timer-12G,Δ%,0.108,
+efficiency,PEMS03,Timer-12G,Δ%,0.135,
+efficiency,PEMS03,Timer-12G,Δ%,0.179,
+efficiency,PEMS03,Timer-12G,Δ%,0.248,
+efficiency,PEMS04,Timer-12G,Δ%,0.134,
+efficiency,PEMS04,Timer-12G,Δ%,0.166,
+efficiency,PEMS04,Timer-12G,Δ%,0.216,
+efficiency,PEMS04,Timer-12G,Δ%,0.296,
+efficiency,PEMS07,Timer-12G,Δ%,0.114,
+efficiency,PEMS07,Timer-12G,Δ%,0.144,
+efficiency,PEMS07,Timer-12G,Δ%,0.189,
+efficiency,PEMS07,Timer-12G,Δ%,0.256,
+efficiency,PEMS08,Timer-12G,Δ%,0.129,
+efficiency,PEMS08,Timer-12G,Δ%,0.157,
+efficiency,PEMS08,Timer-12G,Δ%,0.206,
+efficiency,PEMS08,Timer-12G,Δ%,0.288,
diff --git a/result/per_paper/2402.02368/accuracy_efficiency_traced.csv b/result/per_paper/2402.02368/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..fad1e97d54ba6dedc3c07aa72e40cc7c4283a0e8
--- /dev/null
+++ b/result/per_paper/2402.02368/accuracy_efficiency_traced.csv
@@ -0,0 +1,108 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,,1,1,2
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,,1,1,4
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,,1,1,6
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,,1,1,8
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,,1,1,10
+accuracy,PRE-TRAINED,Timer-12G,MAE,12,,1,1,12
+accuracy,SOTA,Timer-12G,MAE,0.129,,1,14,2
+accuracy,SOTA,Timer-12G,MAE,0.149,,1,14,4
+accuracy,SOTA,Timer-12G,MAE,0.115,,1,14,6
+efficiency,MASK RATIO,Timer-12G,Δ%,25.0,,4,0,2
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,4,1,2
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,4,1,5
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,4,1,8
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,4,1,11
+efficiency,ETTH1,Timer-12G,Δ%,0.278,,4,2,2
+efficiency,ETTH1,Timer-12G,Δ%,0.287,,4,2,5
+efficiency,ETTH1,Timer-12G,Δ%,0.297,,4,2,8
+efficiency,ETTH1,Timer-12G,Δ%,0.314,,4,2,11
+efficiency,ETTH2,Timer-12G,Δ%,0.166,,4,3,2
+efficiency,ETTH2,Timer-12G,Δ%,0.178,,4,3,5
+efficiency,ETTH2,Timer-12G,Δ%,0.190,,4,3,8
+efficiency,ETTH2,Timer-12G,Δ%,0.208,,4,3,11
+efficiency,ETTM1,Timer-12G,Δ%,0.328,,4,4,2
+efficiency,ETTM1,Timer-12G,Δ%,0.326,,4,4,5
+efficiency,ETTM1,Timer-12G,Δ%,0.360,,4,4,8
+efficiency,ETTM1,Timer-12G,Δ%,0.407,,4,4,11
+efficiency,ETTM2,Timer-12G,Δ%,0.133,,4,5,2
+efficiency,ETTM2,Timer-12G,Δ%,0.123,,4,5,5
+efficiency,ETTM2,Timer-12G,Δ%,0.136,,4,5,8
+efficiency,ETTM2,Timer-12G,Δ%,0.143,,4,5,11
+efficiency,ECL,Timer-12G,Δ%,0.130,,4,6,2
+efficiency,ECL,Timer-12G,Δ%,0.138,,4,6,5
+efficiency,ECL,Timer-12G,Δ%,0.149,,4,6,8
+efficiency,ECL,Timer-12G,Δ%,0.162,,4,6,11
+efficiency,TRAFFIC,Timer-12G,Δ%,0.420,,4,7,2
+efficiency,TRAFFIC,Timer-12G,Δ%,0.446,,4,7,5
+efficiency,TRAFFIC,Timer-12G,Δ%,0.477,,4,7,8
+efficiency,TRAFFIC,Timer-12G,Δ%,0.511,,4,7,11
+efficiency,WEATHER,Timer-12G,Δ%,0.129,,4,8,2
+efficiency,WEATHER,Timer-12G,Δ%,0.147,,4,8,5
+efficiency,WEATHER,Timer-12G,Δ%,0.125,,4,8,8
+efficiency,WEATHER,Timer-12G,Δ%,0.153,,4,8,11
+efficiency,PEMS03,Timer-12G,Δ%,0.120,,4,9,2
+efficiency,PEMS03,Timer-12G,Δ%,0.150,,4,9,5
+efficiency,PEMS03,Timer-12G,Δ%,0.198,,4,9,8
+efficiency,PEMS03,Timer-12G,Δ%,0.273,,4,9,11
+efficiency,PEMS04,Timer-12G,Δ%,0.146,,4,10,2
+efficiency,PEMS04,Timer-12G,Δ%,0.184,,4,10,5
+efficiency,PEMS04,Timer-12G,Δ%,0.236,,4,10,8
+efficiency,PEMS04,Timer-12G,Δ%,0.320,,4,10,11
+efficiency,PEMS07,Timer-12G,Δ%,0.125,,4,11,2
+efficiency,PEMS07,Timer-12G,Δ%,0.162,,4,11,5
+efficiency,PEMS07,Timer-12G,Δ%,0.214,,4,11,8
+efficiency,PEMS07,Timer-12G,Δ%,0.290,,4,11,11
+efficiency,PEMS08,Timer-12G,Δ%,0.139,,4,12,2
+efficiency,PEMS08,Timer-12G,Δ%,0.174,,4,12,5
+efficiency,PEMS08,Timer-12G,Δ%,0.236,,4,12,8
+efficiency,PEMS08,Timer-12G,Δ%,0.324,,4,12,11
+efficiency,MASK RATIO,Timer-12G,Δ%,25.0,,5,0,2
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,5,1,2
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,5,1,5
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,5,1,8
+efficiency,PRE-TRAINED,Timer-12G,Δ%,12,,5,1,11
+efficiency,ETTH1,Timer-12G,Δ%,0.273,,5,2,2
+efficiency,ETTH1,Timer-12G,Δ%,0.283,,5,2,5
+efficiency,ETTH1,Timer-12G,Δ%,0.294,,5,2,8
+efficiency,ETTH1,Timer-12G,Δ%,0.312,,5,2,11
+efficiency,ETTH2,Timer-12G,Δ%,0.177,,5,3,2
+efficiency,ETTH2,Timer-12G,Δ%,0.186,,5,3,5
+efficiency,ETTH2,Timer-12G,Δ%,0.195,,5,3,8
+efficiency,ETTH2,Timer-12G,Δ%,0.209,,5,3,11
+efficiency,ETTM1,Timer-12G,Δ%,0.352,,5,4,2
+efficiency,ETTM1,Timer-12G,Δ%,0.345,,5,4,5
+efficiency,ETTM1,Timer-12G,Δ%,0.371,,5,4,8
+efficiency,ETTM1,Timer-12G,Δ%,0.413,,5,4,11
+efficiency,ETTM2,Timer-12G,Δ%,0.161,,5,5,2
+efficiency,ETTM2,Timer-12G,Δ%,0.171,,5,5,5
+efficiency,ETTM2,Timer-12G,Δ%,0.176,,5,5,8
+efficiency,ETTM2,Timer-12G,Δ%,0.158,,5,5,11
+efficiency,ECL,Timer-12G,Δ%,0.122,,5,6,2
+efficiency,ECL,Timer-12G,Δ%,0.130,,5,6,5
+efficiency,ECL,Timer-12G,Δ%,0.139,,5,6,8
+efficiency,ECL,Timer-12G,Δ%,0.152,,5,6,11
+efficiency,TRAFFIC,Timer-12G,Δ%,0.392,,5,7,2
+efficiency,TRAFFIC,Timer-12G,Δ%,0.414,,5,7,5
+efficiency,TRAFFIC,Timer-12G,Δ%,0.443,,5,7,8
+efficiency,TRAFFIC,Timer-12G,Δ%,0.477,,5,7,11
+efficiency,WEATHER,Timer-12G,Δ%,0.157,,5,8,2
+efficiency,WEATHER,Timer-12G,Δ%,0.146,,5,8,5
+efficiency,WEATHER,Timer-12G,Δ%,0.147,,5,8,8
+efficiency,WEATHER,Timer-12G,Δ%,0.158,,5,8,11
+efficiency,PEMS03,Timer-12G,Δ%,0.108,,5,9,2
+efficiency,PEMS03,Timer-12G,Δ%,0.135,,5,9,5
+efficiency,PEMS03,Timer-12G,Δ%,0.179,,5,9,8
+efficiency,PEMS03,Timer-12G,Δ%,0.248,,5,9,11
+efficiency,PEMS04,Timer-12G,Δ%,0.134,,5,10,2
+efficiency,PEMS04,Timer-12G,Δ%,0.166,,5,10,5
+efficiency,PEMS04,Timer-12G,Δ%,0.216,,5,10,8
+efficiency,PEMS04,Timer-12G,Δ%,0.296,,5,10,11
+efficiency,PEMS07,Timer-12G,Δ%,0.114,,5,11,2
+efficiency,PEMS07,Timer-12G,Δ%,0.144,,5,11,5
+efficiency,PEMS07,Timer-12G,Δ%,0.189,,5,11,8
+efficiency,PEMS07,Timer-12G,Δ%,0.256,,5,11,11
+efficiency,PEMS08,Timer-12G,Δ%,0.129,,5,12,2
+efficiency,PEMS08,Timer-12G,Δ%,0.157,,5,12,5
+efficiency,PEMS08,Timer-12G,Δ%,0.206,,5,12,8
+efficiency,PEMS08,Timer-12G,Δ%,0.288,,5,12,11
diff --git a/result/per_paper/2402.02368/components_architecture.csv b/result/per_paper/2402.02368/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a0d2c5600ec50d9f6beb7bf8a5588b5b9a56c547
--- /dev/null
+++ b/result/per_paper/2402.02368/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+GPT-style architecture,"A transformer-based architecture inspired by large language models, designed for autoregressive generation and flexible context length.",reused_cited,"Bengio et al., 2000","we adopt the GPT-style objective that predicts the next token (Bengio et al., 2000)"
+Single-series sequence (S3) format,A data format that converts heterogeneous multivariate time series into unified token sequences for pre-training.,proposed_here,,we propose the single-series sequence (S3) format that converts multivariate series with reserved patterns into unified token sequences
+Time Series Transformer (Timer),"A large-scale pre-trained model for time series analysis, adapted to forecasting, imputation, and anomaly detection.",proposed_here,,"we present Timer, a large-scale pre-trained Time Series Transformer"
+Next token prediction objective,"A pre-training task where the model predicts the next token in a sequence, similar to language models.",reused_cited,"Bengio et al., 2000","we adopt the GPT-style objective that predicts the next token (Bengio et al., 2000)"
+Autoregressive generation,"A generation mechanism where the model predicts subsequent tokens based on prior context, enabling multi-step forecasting.",reused_cited,"Bengio et al., 2000",Timer exhibits similar characteristics as large language models such as... autoregressive generation
+Flexible context length,A design feature allowing the model to handle variable-length input sequences during pre-training and adaptation.,reused_cited,"Bengio et al., 2000",Timer exhibits similar characteristics as large language models such as... flexible context length
diff --git a/result/per_paper/2402.02368/computational.csv b/result/per_paper/2402.02368/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..966083f3078c2a587bd318c7f10fad66862fb7ee
--- /dev/null
+++ b/result/per_paper/2402.02368/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,"29M, 50M, 67M",M,stated,Compute table under 'METHOD | TIMER(OURS)'
+context_lengths_evaluated,1440,tokens,stated,Compute table under 'METHOD | TIMER(OURS)'
+horizon_lengths_evaluated,96,time points,stated,Setup text: 'forecast length as 96'
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2402.02592/accuracy_efficiency.csv b/result/per_paper/2402.02592/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d2dd63125c799372ed3626554bd50782c1970606
--- /dev/null
+++ b/result/per_paper/2402.02592/accuracy_efficiency.csv
@@ -0,0 +1,289 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,Moirai Small,MSE,0.419,
+accuracy,ETTh1,Moirai Base,MSE,0.423,
+accuracy,ETTh1,Moirai Large,MSE,0.448,
+accuracy,ETTh2,Moirai Small,MSE,0.348,
+accuracy,ETTh2,Moirai Base,MSE,0.745,
+accuracy,ETTh2,Moirai Large,MSE,0.584,
+accuracy,ETTm1,Moirai Small,MSE,0.367,
+accuracy,ETTm1,Moirai Base,MSE,0.404,
+accuracy,ETTm1,Moirai Large,MSE,0.426,
+accuracy,ETTm2,Moirai Small,MSE,0.259,
+accuracy,ETTm2,Moirai Base,MSE,0.287,
+accuracy,ETTm2,Moirai Large,MSE,0.366,
+accuracy,Electricity,Moirai Small,MSE,0.285,
+accuracy,Electricity,Moirai Base,MSE,0.219,
+accuracy,Electricity,Moirai Large,MSE,0.314,
+accuracy,Weather,Moirai Small,MSE,0.218,
+accuracy,Weather,Moirai Base,MSE,0.158,
+accuracy,Weather,Moirai Large,MSE,0.230,
+accuracy,M1 Monthly,Moirai Small,MAE,2,
+accuracy,M1 Monthly,Moirai Base,MAE,2,
+accuracy,M1 Monthly,Moirai Large,MAE,1,
+accuracy,M3 Monthly,Moirai Small,MAE,713.41,
+accuracy,M3 Monthly,Moirai Base,MAE,658.17,
+accuracy,M3 Monthly,Moirai Large,MAE,664.03,
+accuracy,M3 Other,Moirai Small,MAE,263.54,
+accuracy,M3 Other,Moirai Base,MAE,198.62,
+accuracy,M3 Other,Moirai Large,MAE,202.41,
+accuracy,M4 Monthly,Moirai Small,MAE,597.6,
+accuracy,M4 Monthly,Moirai Base,MAE,592.09,
+accuracy,M4 Monthly,Moirai Large,MAE,584.36,
+accuracy,M4 Weekly,Moirai Small,MAE,339.76,
+accuracy,M4 Weekly,Moirai Base,MAE,328.08,
+accuracy,M4 Weekly,Moirai Large,MAE,301.52,
+accuracy,M4 Daily,Moirai Small,MAE,189.1,
+accuracy,M4 Daily,Moirai Base,MAE,192.66,
+accuracy,M4 Daily,Moirai Large,MAE,189.78,
+accuracy,M4 Hourly,Moirai Small,MAE,268.04,
+accuracy,M4 Hourly,Moirai Base,MAE,209.87,
+accuracy,M4 Hourly,Moirai Large,MAE,197.79,
+accuracy,Tourism Quarterly,Moirai Small,MAE,18,
+accuracy,Tourism Quarterly,Moirai Base,MAE,17,
+accuracy,Tourism Quarterly,Moirai Large,MAE,15,
+accuracy,Tourism Monthly,Moirai Small,MAE,3,
+accuracy,Tourism Monthly,Moirai Base,MAE,2,
+accuracy,Tourism Monthly,Moirai Large,MAE,2,
+accuracy,CIF 2016,Moirai Small,MAE,655,
+accuracy,CIF 2016,Moirai Base,MAE,539,
+accuracy,CIF 2016,Moirai Large,MAE,695,
+accuracy,Aus. Elec. Demand,Moirai Small,MAE,266.57,
+accuracy,Aus. Elec. Demand,Moirai Base,MAE,201.39,
+accuracy,Aus. Elec. Demand,Moirai Large,MAE,177.68,
+accuracy,Bitcoin,Moirai Small,MAE,1.76,
+accuracy,Bitcoin,Moirai Base,MAE,1.62,
+accuracy,Bitcoin,Moirai Large,MAE,1.87,
+accuracy,Pedestrian Counts,Moirai Small,MAE,54.88,
+accuracy,Pedestrian Counts,Moirai Base,MAE,54.08,
+accuracy,Pedestrian Counts,Moirai Large,MAE,41.66,
+accuracy,Vehicle Trips,Moirai Small,MAE,24.46,
+accuracy,Vehicle Trips,Moirai Base,MAE,23.17,
+accuracy,Vehicle Trips,Moirai Large,MAE,21.85,
+accuracy,KDD cup,Moirai Small,MAE,39.81,
+accuracy,KDD cup,Moirai Base,MAE,38.66,
+accuracy,KDD cup,Moirai Large,MAE,39.09,
+accuracy,Weather,Moirai Small,MAE,1.96,
+accuracy,Weather,Moirai Base,MAE,1.8,
+accuracy,Weather,Moirai Large,MAE,1.75,
+accuracy,NN5 Daily,Moirai Small,MAE,5.37,
+accuracy,NN5 Daily,Moirai Base,MAE,4.26,
+accuracy,NN5 Daily,Moirai Large,MAE,3.77,
+accuracy,NN5 Weekly,Moirai Small,MAE,15.07,
+accuracy,NN5 Weekly,Moirai Base,MAE,16.42,
+accuracy,NN5 Weekly,Moirai Large,MAE,15.3,
+accuracy,Carparts,Moirai Small,MAE,0.53,
+accuracy,Carparts,Moirai Base,MAE,0.47,
+accuracy,Carparts,Moirai Large,MAE,0.49,
+accuracy,FRED-MD,Moirai Small,MAE,2,
+accuracy,FRED-MD,Moirai Base,MAE,2,
+accuracy,FRED-MD,Moirai Large,MAE,2,
+accuracy,Traffic Hourly,Moirai Small,MAE,0.02,
+accuracy,Traffic Hourly,Moirai Base,MAE,0.02,
+accuracy,Traffic Hourly,Moirai Large,MAE,0.01,
+accuracy,Traffic Weekly,Moirai Small,MAE,1.17,
+accuracy,Traffic Weekly,Moirai Base,MAE,1.14,
+accuracy,Traffic Weekly,Moirai Large,MAE,1.13,
+accuracy,Rideshare,Moirai Small,MAE,1.35,
+accuracy,Rideshare,Moirai Base,MAE,1.39,
+accuracy,Rideshare,Moirai Large,MAE,1.29,
+accuracy,Hospital,Moirai Small,MAE,23,
+accuracy,Hospital,Moirai Base,MAE,19.4,
+accuracy,Hospital,Moirai Large,MAE,19.44,
+accuracy,COVID Deaths,Moirai Small,MAE,124.32,
+accuracy,COVID Deaths,Moirai Base,MAE,126.11,
+accuracy,COVID Deaths,Moirai Large,MAE,117.11,
+accuracy,Temperature Rain,Moirai Small,MAE,5.3,
+accuracy,Temperature Rain,Moirai Base,MAE,5.08,
+accuracy,Temperature Rain,Moirai Large,MAE,5.27,
+accuracy,Sunspot,Moirai Small,MAE,0.11,
+accuracy,Sunspot,Moirai Base,MAE,0.08,
+accuracy,Sunspot,Moirai Large,MAE,0.13,
+accuracy,Saugeen River Flow,Moirai Small,MAE,24.07,
+accuracy,Saugeen River Flow,Moirai Base,MAE,24.4,
+accuracy,Saugeen River Flow,Moirai Large,MAE,24.76,
+accuracy,US Births,Moirai Small,MAE,872.51,
+accuracy,US Births,Moirai Base,MAE,624.3,
+accuracy,US Births,Moirai Large,MAE,476.5,
+accuracy,Electricity,MoiraiSmall,CRPS,0.072,
+accuracy,Electricity,MoiraiBase,CRPS,0.055,
+accuracy,Electricity,MoiraiLarge,CRPS,0.050,
+accuracy,MSIS,MoiraiSmall,CRPS,6.172,
+accuracy,MSIS,MoiraiBase,CRPS,5.875,
+accuracy,MSIS,MoiraiLarge,CRPS,5.744,
+accuracy,sMAPE,MoiraiSmall,CRPS,0.111,
+accuracy,sMAPE,MoiraiBase,CRPS,0.106,
+accuracy,sMAPE,MoiraiLarge,CRPS,0.107,
+accuracy,MASE,MoiraiSmall,CRPS,0.792,
+accuracy,MASE,MoiraiBase,CRPS,0.751,
+accuracy,MASE,MoiraiLarge,CRPS,0.753,
+accuracy,ND,MoiraiSmall,CRPS,0.069,
+accuracy,ND,MoiraiBase,CRPS,0.063,
+accuracy,ND,MoiraiLarge,CRPS,0.065,
+accuracy,NRMSE,MoiraiSmall,CRPS,0.551,
+accuracy,NRMSE,MoiraiBase,CRPS,0.465,
+accuracy,NRMSE,MoiraiLarge,CRPS,0.506,
+accuracy,Solar,MoiraiSmall,CRPS,0.471,
+accuracy,Solar,MoiraiBase,CRPS,0.419,
+accuracy,Solar,MoiraiLarge,CRPS,0.406,
+accuracy,MSIS,MoiraiSmall,CRPS,7.011,
+accuracy,MSIS,MoiraiBase,CRPS,6.250,
+accuracy,MSIS,MoiraiLarge,CRPS,8.447,
+accuracy,sMAPE,MoiraiSmall,CRPS,1.410,
+accuracy,sMAPE,MoiraiBase,CRPS,1.400,
+accuracy,sMAPE,MoiraiLarge,CRPS,1.501,
+accuracy,MASE,MoiraiSmall,CRPS,1.292,
+accuracy,MASE,MoiraiBase,CRPS,1.237,
+accuracy,MASE,MoiraiLarge,CRPS,1.607,
+accuracy,ND,MoiraiSmall,CRPS,0.551,
+accuracy,ND,MoiraiBase,CRPS,0.528,
+accuracy,ND,MoiraiLarge,CRPS,0.685,
+accuracy,NRMSE,MoiraiSmall,CRPS,1.034,
+accuracy,NRMSE,MoiraiBase,CRPS,1.014,
+accuracy,NRMSE,MoiraiLarge,CRPS,1.408,
+accuracy,Walmart,MoiraiSmall,CRPS,0.103,
+accuracy,Walmart,MoiraiBase,CRPS,0.093,
+accuracy,Walmart,MoiraiLarge,CRPS,0.098,
+accuracy,MSIS,MoiraiSmall,CRPS,8.421,
+accuracy,MSIS,MoiraiBase,CRPS,8.520,
+accuracy,MSIS,MoiraiLarge,CRPS,6.005,
+accuracy,sMAPE,MoiraiSmall,CRPS,0.168,
+accuracy,sMAPE,MoiraiBase,CRPS,0.174,
+accuracy,sMAPE,MoiraiLarge,CRPS,0.150,
+accuracy,MASE,MoiraiSmall,CRPS,0.964,
+accuracy,MASE,MoiraiBase,CRPS,1.007,
+accuracy,MASE,MoiraiLarge,CRPS,0.867,
+accuracy,ND,MoiraiSmall,CRPS,0.117,
+accuracy,ND,MoiraiBase,CRPS,0.124,
+accuracy,ND,MoiraiLarge,CRPS,0.105,
+accuracy,NRMSE,MoiraiSmall,CRPS,0.291,
+accuracy,NRMSE,MoiraiBase,CRPS,0.332,
+accuracy,NRMSE,MoiraiLarge,CRPS,0.218,
+accuracy,Weather,MoiraiSmall,CRPS,0.049,
+accuracy,Weather,MoiraiBase,CRPS,0.041,
+accuracy,Weather,MoiraiLarge,CRPS,0.051,
+accuracy,MSIS,MoiraiSmall,CRPS,5.136,
+accuracy,MSIS,MoiraiBase,CRPS,4.962,
+accuracy,MSIS,MoiraiLarge,CRPS,7.759,
+accuracy,sMAPE,MoiraiSmall,CRPS,0.623,
+accuracy,sMAPE,MoiraiBase,CRPS,0.688,
+accuracy,sMAPE,MoiraiLarge,CRPS,0.668,
+accuracy,MASE,MoiraiSmall,CRPS,0.487,
+accuracy,MASE,MoiraiBase,CRPS,0.515,
+accuracy,MASE,MoiraiLarge,CRPS,0.844,
+accuracy,ND,MoiraiSmall,CRPS,0.048,
+accuracy,ND,MoiraiBase,CRPS,0.063,
+accuracy,ND,MoiraiLarge,CRPS,0.072,
+accuracy,NRMSE,MoiraiSmall,CRPS,0.417,
+accuracy,NRMSE,MoiraiBase,CRPS,0.331,
+accuracy,NRMSE,MoiraiLarge,CRPS,0.260,
+accuracy,Istanbul Traffic,MoiraiSmall,CRPS,0.173,
+accuracy,Istanbul Traffic,MoiraiBase,CRPS,0.116,
+accuracy,Istanbul Traffic,MoiraiLarge,CRPS,0.112,
+accuracy,MSIS,MoiraiSmall,CRPS,4.461,
+accuracy,MSIS,MoiraiBase,CRPS,4.277,
+accuracy,MSIS,MoiraiLarge,CRPS,3.813,
+accuracy,sMAPE,MoiraiSmall,CRPS,0.284,
+accuracy,sMAPE,MoiraiBase,CRPS,0.288,
+accuracy,sMAPE,MoiraiLarge,CRPS,0.287,
+accuracy,MASE,MoiraiSmall,CRPS,0.644,
+accuracy,MASE,MoiraiBase,CRPS,0.631,
+accuracy,MASE,MoiraiLarge,CRPS,0.653,
+accuracy,ND,MoiraiSmall,CRPS,0.146,
+accuracy,ND,MoiraiBase,CRPS,0.143,
+accuracy,ND,MoiraiLarge,CRPS,0.148,
+accuracy,NRMSE,MoiraiSmall,CRPS,0.194,
+accuracy,NRMSE,MoiraiBase,CRPS,0.186,
+accuracy,NRMSE,MoiraiLarge,CRPS,0.190,
+accuracy,Turkey Power,MoiraiSmall,CRPS,0.048,
+accuracy,Turkey Power,MoiraiBase,CRPS,0.040,
+accuracy,Turkey Power,MoiraiLarge,CRPS,0.036,
+accuracy,MSIS,MoiraiSmall,CRPS,6.766,
+accuracy,MSIS,MoiraiBase,CRPS,6.341,
+accuracy,MSIS,MoiraiLarge,CRPS,8.978,
+accuracy,sMAPE,MoiraiSmall,CRPS,0.378,
+accuracy,sMAPE,MoiraiBase,CRPS,0.375,
+accuracy,sMAPE,MoiraiLarge,CRPS,0.416,
+accuracy,MASE,MoiraiSmall,CRPS,0.888,
+accuracy,MASE,MoiraiBase,CRPS,0.870,
+accuracy,MASE,MoiraiLarge,CRPS,1.234,
+accuracy,ND,MoiraiSmall,CRPS,0.051,
+accuracy,ND,MoiraiBase,CRPS,0.046,
+accuracy,ND,MoiraiLarge,CRPS,0.071,
+accuracy,NRMSE,MoiraiSmall,CRPS,0.118,
+accuracy,NRMSE,MoiraiBase,CRPS,0.102,
+accuracy,NRMSE,MoiraiLarge,CRPS,0.158,
+accuracy,Electricity,Moirai Small,CRPS,0.072,
+accuracy,Electricity,Moirai Base,CRPS,0.055,
+accuracy,Electricity,Moirai Large,CRPS,0.050,
+accuracy,MSIS,Moirai Small,CRPS,6.172,
+accuracy,MSIS,Moirai Base,CRPS,5.875,
+accuracy,MSIS,Moirai Large,CRPS,5.744,0.12
+accuracy,Solar,Moirai Small,CRPS,0.471,
+accuracy,Solar,Moirai Base,CRPS,0.419,
+accuracy,Solar,Moirai Large,CRPS,0.406,
+accuracy,MSIS,Moirai Small,CRPS,7.011,
+accuracy,MSIS,Moirai Base,CRPS,6.250,
+accuracy,MSIS,Moirai Large,CRPS,8.447,1.59
+accuracy,Walmart,Moirai Small,CRPS,0.103,
+accuracy,Walmart,Moirai Base,CRPS,0.093,
+accuracy,Walmart,Moirai Large,CRPS,0.098,
+accuracy,MSIS,Moirai Small,CRPS,8.421,
+accuracy,MSIS,Moirai Base,CRPS,8.520,
+accuracy,MSIS,Moirai Large,CRPS,6.005,0.21
+accuracy,Weather,Moirai Small,CRPS,0.049,
+accuracy,Weather,Moirai Base,CRPS,0.041,
+accuracy,Weather,Moirai Large,CRPS,0.051,
+accuracy,MSIS,Moirai Small,CRPS,5.136,
+accuracy,MSIS,Moirai Base,CRPS,4.962,
+accuracy,MSIS,Moirai Large,CRPS,7.759,0.49
+accuracy,Istanbul Traffic,Moirai Small,CRPS,0.173,
+accuracy,Istanbul Traffic,Moirai Base,CRPS,0.116,
+accuracy,Istanbul Traffic,Moirai Large,CRPS,0.112,
+accuracy,MSIS,Moirai Small,CRPS,4.461,
+accuracy,MSIS,Moirai Base,CRPS,4.277,
+accuracy,MSIS,Moirai Large,CRPS,3.813,0.09
+accuracy,Turkey Power,Moirai Small,CRPS,0.048,
+accuracy,Turkey Power,Moirai Base,CRPS,0.040,
+accuracy,Turkey Power,Moirai Large,CRPS,0.036,
+accuracy,MSIS,Moirai Small,CRPS,6.766,
+accuracy,MSIS,Moirai Base,CRPS,6.341,
+accuracy,MSIS,Moirai Large,CRPS,8.978,0.51
+accuracy,ETTh1,Moirai Small,MSE,0.400,
+accuracy,ETTh1,Moirai Base,MSE,0.434,
+accuracy,ETTh1,Moirai Large,MSE,0.510,
+accuracy,MAE,Moirai Small,MSE,0.438,
+accuracy,MAE,Moirai Base,MSE,0.469,
+accuracy,MAE,Moirai Large,MSE,0.448,
+accuracy,ETTh2,Moirai Small,MSE,0.341,
+accuracy,ETTh2,Moirai Base,MSE,0.345,
+accuracy,ETTh2,Moirai Large,MSE,0.354,
+accuracy,MAE,Moirai Small,MSE,0.382,
+accuracy,MAE,Moirai Base,MSE,0.376,
+accuracy,MAE,Moirai Large,MSE,0.407,
+accuracy,ETTm1,Moirai Small,MSE,0.448,
+accuracy,ETTm1,Moirai Base,MSE,0.381,
+accuracy,ETTm1,Moirai Large,MSE,0.390,
+accuracy,MAE,Moirai Small,MSE,0.388,
+accuracy,MAE,Moirai Base,MSE,0.389,
+accuracy,MAE,Moirai Large,MSE,0.410,
+accuracy,ETTm2,Moirai Small,MSE,0.300,
+accuracy,ETTm2,Moirai Base,MSE,0.272,
+accuracy,ETTm2,Moirai Large,MSE,0.276,
+accuracy,MAE,Moirai Small,MSE,0.321,
+accuracy,MAE,Moirai Base,MSE,0.320,
+accuracy,MAE,Moirai Large,MSE,0.332,
+accuracy,Electricity,Moirai Small,MSE,0.233,
+accuracy,Electricity,Moirai Base,MSE,0.188,
+accuracy,Electricity,Moirai Large,MSE,0.188,
+accuracy,MAE,Moirai Small,MSE,0.274,
+accuracy,MAE,Moirai Base,MSE,0.273,
+accuracy,MAE,Moirai Large,MSE,0.270,
+accuracy,Weather,Moirai Small,MSE,0.242,
+accuracy,Weather,Moirai Base,MSE,0.238,
+accuracy,Weather,Moirai Large,MSE,0.259,
+accuracy,MAE,Moirai Small,MSE,0.261,
+accuracy,MAE,Moirai Base,MSE,0.275,
+accuracy,MAE,Moirai Large,MSE,0.278,
+efficiency,$MOIRAI_{Small} (32)$,Moirai-Small,seconds,32,
+efficiency,$MOIRAI_{Base} (32)$,Moirai-Small,seconds,32,
+efficiency,$MOIRAI_{Large} (32)$,Moirai-Small,seconds,32,
diff --git a/result/per_paper/2402.02592/accuracy_efficiency_traced.csv b/result/per_paper/2402.02592/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..64f969c69d616fcb68df506392153a9d78ae495e
--- /dev/null
+++ b/result/per_paper/2402.02592/accuracy_efficiency_traced.csv
@@ -0,0 +1,289 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,Moirai Small,MSE,0.419,,0,3,13
+accuracy,ETTh1,Moirai Base,MSE,0.423,,0,3,14
+accuracy,ETTh1,Moirai Large,MSE,0.448,,0,3,15
+accuracy,ETTh2,Moirai Small,MSE,0.348,,0,7,13
+accuracy,ETTh2,Moirai Base,MSE,0.745,,0,7,14
+accuracy,ETTh2,Moirai Large,MSE,0.584,,0,7,15
+accuracy,ETTm1,Moirai Small,MSE,0.367,,0,11,13
+accuracy,ETTm1,Moirai Base,MSE,0.404,,0,11,14
+accuracy,ETTm1,Moirai Large,MSE,0.426,,0,11,15
+accuracy,ETTm2,Moirai Small,MSE,0.259,,0,15,13
+accuracy,ETTm2,Moirai Base,MSE,0.287,,0,15,14
+accuracy,ETTm2,Moirai Large,MSE,0.366,,0,15,15
+accuracy,Electricity,Moirai Small,MSE,0.285,,0,19,13
+accuracy,Electricity,Moirai Base,MSE,0.219,,0,19,14
+accuracy,Electricity,Moirai Large,MSE,0.314,,0,19,15
+accuracy,Weather,Moirai Small,MSE,0.218,,0,23,13
+accuracy,Weather,Moirai Base,MSE,0.158,,0,23,14
+accuracy,Weather,Moirai Large,MSE,0.230,,0,23,15
+accuracy,M1 Monthly,Moirai Small,MAE,2,,1,1,1
+accuracy,M1 Monthly,Moirai Base,MAE,2,,1,1,2
+accuracy,M1 Monthly,Moirai Large,MAE,1,,1,1,3
+accuracy,M3 Monthly,Moirai Small,MAE,713.41,,1,2,1
+accuracy,M3 Monthly,Moirai Base,MAE,658.17,,1,2,2
+accuracy,M3 Monthly,Moirai Large,MAE,664.03,,1,2,3
+accuracy,M3 Other,Moirai Small,MAE,263.54,,1,3,1
+accuracy,M3 Other,Moirai Base,MAE,198.62,,1,3,2
+accuracy,M3 Other,Moirai Large,MAE,202.41,,1,3,3
+accuracy,M4 Monthly,Moirai Small,MAE,597.6,,1,4,1
+accuracy,M4 Monthly,Moirai Base,MAE,592.09,,1,4,2
+accuracy,M4 Monthly,Moirai Large,MAE,584.36,,1,4,3
+accuracy,M4 Weekly,Moirai Small,MAE,339.76,,1,5,1
+accuracy,M4 Weekly,Moirai Base,MAE,328.08,,1,5,2
+accuracy,M4 Weekly,Moirai Large,MAE,301.52,,1,5,3
+accuracy,M4 Daily,Moirai Small,MAE,189.1,,1,6,1
+accuracy,M4 Daily,Moirai Base,MAE,192.66,,1,6,2
+accuracy,M4 Daily,Moirai Large,MAE,189.78,,1,6,3
+accuracy,M4 Hourly,Moirai Small,MAE,268.04,,1,7,1
+accuracy,M4 Hourly,Moirai Base,MAE,209.87,,1,7,2
+accuracy,M4 Hourly,Moirai Large,MAE,197.79,,1,7,3
+accuracy,Tourism Quarterly,Moirai Small,MAE,18,,1,8,1
+accuracy,Tourism Quarterly,Moirai Base,MAE,17,,1,8,2
+accuracy,Tourism Quarterly,Moirai Large,MAE,15,,1,8,3
+accuracy,Tourism Monthly,Moirai Small,MAE,3,,1,9,1
+accuracy,Tourism Monthly,Moirai Base,MAE,2,,1,9,2
+accuracy,Tourism Monthly,Moirai Large,MAE,2,,1,9,3
+accuracy,CIF 2016,Moirai Small,MAE,655,,1,10,1
+accuracy,CIF 2016,Moirai Base,MAE,539,,1,10,2
+accuracy,CIF 2016,Moirai Large,MAE,695,,1,10,3
+accuracy,Aus. Elec. Demand,Moirai Small,MAE,266.57,,1,11,1
+accuracy,Aus. Elec. Demand,Moirai Base,MAE,201.39,,1,11,2
+accuracy,Aus. Elec. Demand,Moirai Large,MAE,177.68,,1,11,3
+accuracy,Bitcoin,Moirai Small,MAE,1.76,,1,12,1
+accuracy,Bitcoin,Moirai Base,MAE,1.62,,1,12,2
+accuracy,Bitcoin,Moirai Large,MAE,1.87,,1,12,3
+accuracy,Pedestrian Counts,Moirai Small,MAE,54.88,,1,13,1
+accuracy,Pedestrian Counts,Moirai Base,MAE,54.08,,1,13,2
+accuracy,Pedestrian Counts,Moirai Large,MAE,41.66,,1,13,3
+accuracy,Vehicle Trips,Moirai Small,MAE,24.46,,1,14,1
+accuracy,Vehicle Trips,Moirai Base,MAE,23.17,,1,14,2
+accuracy,Vehicle Trips,Moirai Large,MAE,21.85,,1,14,3
+accuracy,KDD cup,Moirai Small,MAE,39.81,,1,15,1
+accuracy,KDD cup,Moirai Base,MAE,38.66,,1,15,2
+accuracy,KDD cup,Moirai Large,MAE,39.09,,1,15,3
+accuracy,Weather,Moirai Small,MAE,1.96,,1,16,1
+accuracy,Weather,Moirai Base,MAE,1.8,,1,16,2
+accuracy,Weather,Moirai Large,MAE,1.75,,1,16,3
+accuracy,NN5 Daily,Moirai Small,MAE,5.37,,1,17,1
+accuracy,NN5 Daily,Moirai Base,MAE,4.26,,1,17,2
+accuracy,NN5 Daily,Moirai Large,MAE,3.77,,1,17,3
+accuracy,NN5 Weekly,Moirai Small,MAE,15.07,,1,18,1
+accuracy,NN5 Weekly,Moirai Base,MAE,16.42,,1,18,2
+accuracy,NN5 Weekly,Moirai Large,MAE,15.3,,1,18,3
+accuracy,Carparts,Moirai Small,MAE,0.53,,1,19,1
+accuracy,Carparts,Moirai Base,MAE,0.47,,1,19,2
+accuracy,Carparts,Moirai Large,MAE,0.49,,1,19,3
+accuracy,FRED-MD,Moirai Small,MAE,2,,1,20,1
+accuracy,FRED-MD,Moirai Base,MAE,2,,1,20,2
+accuracy,FRED-MD,Moirai Large,MAE,2,,1,20,3
+accuracy,Traffic Hourly,Moirai Small,MAE,0.02,,1,21,1
+accuracy,Traffic Hourly,Moirai Base,MAE,0.02,,1,21,2
+accuracy,Traffic Hourly,Moirai Large,MAE,0.01,,1,21,3
+accuracy,Traffic Weekly,Moirai Small,MAE,1.17,,1,22,1
+accuracy,Traffic Weekly,Moirai Base,MAE,1.14,,1,22,2
+accuracy,Traffic Weekly,Moirai Large,MAE,1.13,,1,22,3
+accuracy,Rideshare,Moirai Small,MAE,1.35,,1,23,1
+accuracy,Rideshare,Moirai Base,MAE,1.39,,1,23,2
+accuracy,Rideshare,Moirai Large,MAE,1.29,,1,23,3
+accuracy,Hospital,Moirai Small,MAE,23,,1,24,1
+accuracy,Hospital,Moirai Base,MAE,19.4,,1,24,2
+accuracy,Hospital,Moirai Large,MAE,19.44,,1,24,3
+accuracy,COVID Deaths,Moirai Small,MAE,124.32,,1,25,1
+accuracy,COVID Deaths,Moirai Base,MAE,126.11,,1,25,2
+accuracy,COVID Deaths,Moirai Large,MAE,117.11,,1,25,3
+accuracy,Temperature Rain,Moirai Small,MAE,5.3,,1,26,1
+accuracy,Temperature Rain,Moirai Base,MAE,5.08,,1,26,2
+accuracy,Temperature Rain,Moirai Large,MAE,5.27,,1,26,3
+accuracy,Sunspot,Moirai Small,MAE,0.11,,1,27,1
+accuracy,Sunspot,Moirai Base,MAE,0.08,,1,27,2
+accuracy,Sunspot,Moirai Large,MAE,0.13,,1,27,3
+accuracy,Saugeen River Flow,Moirai Small,MAE,24.07,,1,28,1
+accuracy,Saugeen River Flow,Moirai Base,MAE,24.4,,1,28,2
+accuracy,Saugeen River Flow,Moirai Large,MAE,24.76,,1,28,3
+accuracy,US Births,Moirai Small,MAE,872.51,,1,29,1
+accuracy,US Births,Moirai Base,MAE,624.3,,1,29,2
+accuracy,US Births,Moirai Large,MAE,476.5,,1,29,3
+accuracy,Electricity,MoiraiSmall,CRPS,0.072,,2,2,2
+accuracy,Electricity,MoiraiBase,CRPS,0.055,,2,2,3
+accuracy,Electricity,MoiraiLarge,CRPS,0.050,,2,2,4
+accuracy,MSIS,MoiraiSmall,CRPS,6.172,,2,3,2
+accuracy,MSIS,MoiraiBase,CRPS,5.875,,2,3,3
+accuracy,MSIS,MoiraiLarge,CRPS,5.744,,2,3,4
+accuracy,sMAPE,MoiraiSmall,CRPS,0.111,,2,4,2
+accuracy,sMAPE,MoiraiBase,CRPS,0.106,,2,4,3
+accuracy,sMAPE,MoiraiLarge,CRPS,0.107,,2,4,4
+accuracy,MASE,MoiraiSmall,CRPS,0.792,,2,5,2
+accuracy,MASE,MoiraiBase,CRPS,0.751,,2,5,3
+accuracy,MASE,MoiraiLarge,CRPS,0.753,,2,5,4
+accuracy,ND,MoiraiSmall,CRPS,0.069,,2,6,2
+accuracy,ND,MoiraiBase,CRPS,0.063,,2,6,3
+accuracy,ND,MoiraiLarge,CRPS,0.065,,2,6,4
+accuracy,NRMSE,MoiraiSmall,CRPS,0.551,,2,7,2
+accuracy,NRMSE,MoiraiBase,CRPS,0.465,,2,7,3
+accuracy,NRMSE,MoiraiLarge,CRPS,0.506,,2,7,4
+accuracy,Solar,MoiraiSmall,CRPS,0.471,,2,8,2
+accuracy,Solar,MoiraiBase,CRPS,0.419,,2,8,3
+accuracy,Solar,MoiraiLarge,CRPS,0.406,,2,8,4
+accuracy,MSIS,MoiraiSmall,CRPS,7.011,,2,9,2
+accuracy,MSIS,MoiraiBase,CRPS,6.250,,2,9,3
+accuracy,MSIS,MoiraiLarge,CRPS,8.447,,2,9,4
+accuracy,sMAPE,MoiraiSmall,CRPS,1.410,,2,10,2
+accuracy,sMAPE,MoiraiBase,CRPS,1.400,,2,10,3
+accuracy,sMAPE,MoiraiLarge,CRPS,1.501,,2,10,4
+accuracy,MASE,MoiraiSmall,CRPS,1.292,,2,11,2
+accuracy,MASE,MoiraiBase,CRPS,1.237,,2,11,3
+accuracy,MASE,MoiraiLarge,CRPS,1.607,,2,11,4
+accuracy,ND,MoiraiSmall,CRPS,0.551,,2,12,2
+accuracy,ND,MoiraiBase,CRPS,0.528,,2,12,3
+accuracy,ND,MoiraiLarge,CRPS,0.685,,2,12,4
+accuracy,NRMSE,MoiraiSmall,CRPS,1.034,,2,13,2
+accuracy,NRMSE,MoiraiBase,CRPS,1.014,,2,13,3
+accuracy,NRMSE,MoiraiLarge,CRPS,1.408,,2,13,4
+accuracy,Walmart,MoiraiSmall,CRPS,0.103,,2,14,2
+accuracy,Walmart,MoiraiBase,CRPS,0.093,,2,14,3
+accuracy,Walmart,MoiraiLarge,CRPS,0.098,,2,14,4
+accuracy,MSIS,MoiraiSmall,CRPS,8.421,,2,15,2
+accuracy,MSIS,MoiraiBase,CRPS,8.520,,2,15,3
+accuracy,MSIS,MoiraiLarge,CRPS,6.005,,2,15,4
+accuracy,sMAPE,MoiraiSmall,CRPS,0.168,,2,16,2
+accuracy,sMAPE,MoiraiBase,CRPS,0.174,,2,16,3
+accuracy,sMAPE,MoiraiLarge,CRPS,0.150,,2,16,4
+accuracy,MASE,MoiraiSmall,CRPS,0.964,,2,17,2
+accuracy,MASE,MoiraiBase,CRPS,1.007,,2,17,3
+accuracy,MASE,MoiraiLarge,CRPS,0.867,,2,17,4
+accuracy,ND,MoiraiSmall,CRPS,0.117,,2,18,2
+accuracy,ND,MoiraiBase,CRPS,0.124,,2,18,3
+accuracy,ND,MoiraiLarge,CRPS,0.105,,2,18,4
+accuracy,NRMSE,MoiraiSmall,CRPS,0.291,,2,19,2
+accuracy,NRMSE,MoiraiBase,CRPS,0.332,,2,19,3
+accuracy,NRMSE,MoiraiLarge,CRPS,0.218,,2,19,4
+accuracy,Weather,MoiraiSmall,CRPS,0.049,,2,20,2
+accuracy,Weather,MoiraiBase,CRPS,0.041,,2,20,3
+accuracy,Weather,MoiraiLarge,CRPS,0.051,,2,20,4
+accuracy,MSIS,MoiraiSmall,CRPS,5.136,,2,21,2
+accuracy,MSIS,MoiraiBase,CRPS,4.962,,2,21,3
+accuracy,MSIS,MoiraiLarge,CRPS,7.759,,2,21,4
+accuracy,sMAPE,MoiraiSmall,CRPS,0.623,,2,22,2
+accuracy,sMAPE,MoiraiBase,CRPS,0.688,,2,22,3
+accuracy,sMAPE,MoiraiLarge,CRPS,0.668,,2,22,4
+accuracy,MASE,MoiraiSmall,CRPS,0.487,,2,23,2
+accuracy,MASE,MoiraiBase,CRPS,0.515,,2,23,3
+accuracy,MASE,MoiraiLarge,CRPS,0.844,,2,23,4
+accuracy,ND,MoiraiSmall,CRPS,0.048,,2,24,2
+accuracy,ND,MoiraiBase,CRPS,0.063,,2,24,3
+accuracy,ND,MoiraiLarge,CRPS,0.072,,2,24,4
+accuracy,NRMSE,MoiraiSmall,CRPS,0.417,,2,25,2
+accuracy,NRMSE,MoiraiBase,CRPS,0.331,,2,25,3
+accuracy,NRMSE,MoiraiLarge,CRPS,0.260,,2,25,4
+accuracy,Istanbul Traffic,MoiraiSmall,CRPS,0.173,,2,26,2
+accuracy,Istanbul Traffic,MoiraiBase,CRPS,0.116,,2,26,3
+accuracy,Istanbul Traffic,MoiraiLarge,CRPS,0.112,,2,26,4
+accuracy,MSIS,MoiraiSmall,CRPS,4.461,,2,27,2
+accuracy,MSIS,MoiraiBase,CRPS,4.277,,2,27,3
+accuracy,MSIS,MoiraiLarge,CRPS,3.813,,2,27,4
+accuracy,sMAPE,MoiraiSmall,CRPS,0.284,,2,28,2
+accuracy,sMAPE,MoiraiBase,CRPS,0.288,,2,28,3
+accuracy,sMAPE,MoiraiLarge,CRPS,0.287,,2,28,4
+accuracy,MASE,MoiraiSmall,CRPS,0.644,,2,29,2
+accuracy,MASE,MoiraiBase,CRPS,0.631,,2,29,3
+accuracy,MASE,MoiraiLarge,CRPS,0.653,,2,29,4
+accuracy,ND,MoiraiSmall,CRPS,0.146,,2,30,2
+accuracy,ND,MoiraiBase,CRPS,0.143,,2,30,3
+accuracy,ND,MoiraiLarge,CRPS,0.148,,2,30,4
+accuracy,NRMSE,MoiraiSmall,CRPS,0.194,,2,31,2
+accuracy,NRMSE,MoiraiBase,CRPS,0.186,,2,31,3
+accuracy,NRMSE,MoiraiLarge,CRPS,0.190,,2,31,4
+accuracy,Turkey Power,MoiraiSmall,CRPS,0.048,,2,32,2
+accuracy,Turkey Power,MoiraiBase,CRPS,0.040,,2,32,3
+accuracy,Turkey Power,MoiraiLarge,CRPS,0.036,,2,32,4
+accuracy,MSIS,MoiraiSmall,CRPS,6.766,,2,33,2
+accuracy,MSIS,MoiraiBase,CRPS,6.341,,2,33,3
+accuracy,MSIS,MoiraiLarge,CRPS,8.978,,2,33,4
+accuracy,sMAPE,MoiraiSmall,CRPS,0.378,,2,34,2
+accuracy,sMAPE,MoiraiBase,CRPS,0.375,,2,34,3
+accuracy,sMAPE,MoiraiLarge,CRPS,0.416,,2,34,4
+accuracy,MASE,MoiraiSmall,CRPS,0.888,,2,35,2
+accuracy,MASE,MoiraiBase,CRPS,0.870,,2,35,3
+accuracy,MASE,MoiraiLarge,CRPS,1.234,,2,35,4
+accuracy,ND,MoiraiSmall,CRPS,0.051,,2,36,2
+accuracy,ND,MoiraiBase,CRPS,0.046,,2,36,3
+accuracy,ND,MoiraiLarge,CRPS,0.071,,2,36,4
+accuracy,NRMSE,MoiraiSmall,CRPS,0.118,,2,37,2
+accuracy,NRMSE,MoiraiBase,CRPS,0.102,,2,37,3
+accuracy,NRMSE,MoiraiLarge,CRPS,0.158,,2,37,4
+accuracy,Electricity,Moirai Small,CRPS,0.072,,3,2,2
+accuracy,Electricity,Moirai Base,CRPS,0.055,,3,2,3
+accuracy,Electricity,Moirai Large,CRPS,0.050,,3,2,4
+accuracy,MSIS,Moirai Small,CRPS,6.172,,3,3,2
+accuracy,MSIS,Moirai Base,CRPS,5.875,,3,3,3
+accuracy,MSIS,Moirai Large,CRPS,5.744,0.12,3,3,4
+accuracy,Solar,Moirai Small,CRPS,0.471,,3,4,2
+accuracy,Solar,Moirai Base,CRPS,0.419,,3,4,3
+accuracy,Solar,Moirai Large,CRPS,0.406,,3,4,4
+accuracy,MSIS,Moirai Small,CRPS,7.011,,3,5,2
+accuracy,MSIS,Moirai Base,CRPS,6.250,,3,5,3
+accuracy,MSIS,Moirai Large,CRPS,8.447,1.59,3,5,4
+accuracy,Walmart,Moirai Small,CRPS,0.103,,3,6,2
+accuracy,Walmart,Moirai Base,CRPS,0.093,,3,6,3
+accuracy,Walmart,Moirai Large,CRPS,0.098,,3,6,4
+accuracy,MSIS,Moirai Small,CRPS,8.421,,3,7,2
+accuracy,MSIS,Moirai Base,CRPS,8.520,,3,7,3
+accuracy,MSIS,Moirai Large,CRPS,6.005,0.21,3,7,4
+accuracy,Weather,Moirai Small,CRPS,0.049,,3,8,2
+accuracy,Weather,Moirai Base,CRPS,0.041,,3,8,3
+accuracy,Weather,Moirai Large,CRPS,0.051,,3,8,4
+accuracy,MSIS,Moirai Small,CRPS,5.136,,3,9,2
+accuracy,MSIS,Moirai Base,CRPS,4.962,,3,9,3
+accuracy,MSIS,Moirai Large,CRPS,7.759,0.49,3,9,4
+accuracy,Istanbul Traffic,Moirai Small,CRPS,0.173,,3,10,2
+accuracy,Istanbul Traffic,Moirai Base,CRPS,0.116,,3,10,3
+accuracy,Istanbul Traffic,Moirai Large,CRPS,0.112,,3,10,4
+accuracy,MSIS,Moirai Small,CRPS,4.461,,3,11,2
+accuracy,MSIS,Moirai Base,CRPS,4.277,,3,11,3
+accuracy,MSIS,Moirai Large,CRPS,3.813,0.09,3,11,4
+accuracy,Turkey Power,Moirai Small,CRPS,0.048,,3,12,2
+accuracy,Turkey Power,Moirai Base,CRPS,0.040,,3,12,3
+accuracy,Turkey Power,Moirai Large,CRPS,0.036,,3,12,4
+accuracy,MSIS,Moirai Small,CRPS,6.766,,3,13,2
+accuracy,MSIS,Moirai Base,CRPS,6.341,,3,13,3
+accuracy,MSIS,Moirai Large,CRPS,8.978,0.51,3,13,4
+accuracy,ETTh1,Moirai Small,MSE,0.400,,4,2,2
+accuracy,ETTh1,Moirai Base,MSE,0.434,,4,2,3
+accuracy,ETTh1,Moirai Large,MSE,0.510,,4,2,4
+accuracy,MAE,Moirai Small,MSE,0.438,,4,3,2
+accuracy,MAE,Moirai Base,MSE,0.469,,4,3,3
+accuracy,MAE,Moirai Large,MSE,0.448,,4,3,4
+accuracy,ETTh2,Moirai Small,MSE,0.341,,4,4,2
+accuracy,ETTh2,Moirai Base,MSE,0.345,,4,4,3
+accuracy,ETTh2,Moirai Large,MSE,0.354,,4,4,4
+accuracy,MAE,Moirai Small,MSE,0.382,,4,5,2
+accuracy,MAE,Moirai Base,MSE,0.376,,4,5,3
+accuracy,MAE,Moirai Large,MSE,0.407,,4,5,4
+accuracy,ETTm1,Moirai Small,MSE,0.448,,4,6,2
+accuracy,ETTm1,Moirai Base,MSE,0.381,,4,6,3
+accuracy,ETTm1,Moirai Large,MSE,0.390,,4,6,4
+accuracy,MAE,Moirai Small,MSE,0.388,,4,7,2
+accuracy,MAE,Moirai Base,MSE,0.389,,4,7,3
+accuracy,MAE,Moirai Large,MSE,0.410,,4,7,4
+accuracy,ETTm2,Moirai Small,MSE,0.300,,4,8,2
+accuracy,ETTm2,Moirai Base,MSE,0.272,,4,8,3
+accuracy,ETTm2,Moirai Large,MSE,0.276,,4,8,4
+accuracy,MAE,Moirai Small,MSE,0.321,,4,9,2
+accuracy,MAE,Moirai Base,MSE,0.320,,4,9,3
+accuracy,MAE,Moirai Large,MSE,0.332,,4,9,4
+accuracy,Electricity,Moirai Small,MSE,0.233,,4,10,2
+accuracy,Electricity,Moirai Base,MSE,0.188,,4,10,3
+accuracy,Electricity,Moirai Large,MSE,0.188,,4,10,4
+accuracy,MAE,Moirai Small,MSE,0.274,,4,11,2
+accuracy,MAE,Moirai Base,MSE,0.273,,4,11,3
+accuracy,MAE,Moirai Large,MSE,0.270,,4,11,4
+accuracy,Weather,Moirai Small,MSE,0.242,,4,12,2
+accuracy,Weather,Moirai Base,MSE,0.238,,4,12,3
+accuracy,Weather,Moirai Large,MSE,0.259,,4,12,4
+accuracy,MAE,Moirai Small,MSE,0.261,,4,13,2
+accuracy,MAE,Moirai Base,MSE,0.275,,4,13,3
+accuracy,MAE,Moirai Large,MSE,0.278,,4,13,4
+efficiency,$MOIRAI_{Small} (32)$,Moirai-Small,seconds,32,,5,2,0
+efficiency,$MOIRAI_{Base} (32)$,Moirai-Small,seconds,32,,5,3,0
+efficiency,$MOIRAI_{Large} (32)$,Moirai-Small,seconds,32,,5,4,0
diff --git a/result/per_paper/2402.02592/components_architecture.csv b/result/per_paper/2402.02592/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a73e9a7c5bcb0a5a5c1ff68ecc82e3399412d4a4
--- /dev/null
+++ b/result/per_paper/2402.02592/components_architecture.csv
@@ -0,0 +1,8 @@
+component,what_it_is,provenance,citation,evidence
+Masked Encoder-based Architecture,"A base architecture for pre-trained time series forecasting models, modified to handle heterogeneity in time series data.",reused_cited,"Woo et al., 2023","Starting from a masked encoder architecture which has been shown to be a strong candidate architecture for scaling up pre-trained time series forecasting models (Woo et al., 2023)"
+Multiple Input/Output Projection Layers,Specialized layers that use patch-based projections with varying patch sizes to handle time series of different frequencies.,proposed_here,,we propose to learn multiple input and output projection layers to handle the differing patterns from time series of varying frequencies. Using patch-based projections with larger patch sizes for high-frequency data and vice versa
+Any-variate Attention,"A mechanism that simultaneously considers time and variate axes as a single sequence, enabling arbitrary multivariate input.",proposed_here,,"our proposed Any-variate Attention, which simultaneously considers both time and variate axes as a single sequence"
+Rotary Position Embeddings (RoPE),Positional encoding technique used to encode time axes in Any-variate Attention.,reused_cited,"Su et al., 2024","leveraging Rotary Position Embeddings (RoPE) (Su et al., 2024)"
+Learned Binary Attention Biases,Attention bias mechanism used to encode variate axes in Any-variate Attention.,reused_cited,"Yang et al., 2022b","learned binary attention biases (Yang et al., 2022b) to encode time and variate axes respectively"
+Mixture of Parametric Distributions,A flexible probabilistic forecasting framework using a combination of parametric distributions for output.,proposed_here,,we overcome the issue of requiring flexible predictive distributions with a mixture of parametric distributions
+Large-scale Open Time Series Archive (LOTSA),A newly introduced dataset with 27B observations across nine domains for pre-training universal forecasting models.,proposed_here,,"we introduce the Large-scale Open Time Series Archive (LOTSA), the largest collection of open time series datasets with 27B observations across nine domains"
diff --git a/result/per_paper/2402.02592/computational.csv b/result/per_paper/2402.02592/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2413d02dff4d8fd884bed6ebadab5075c2570e14
--- /dev/null
+++ b/result/per_paper/2402.02592/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,The paper does not specify the hardware type used for training or inference.
+num_devices,,,not_reported,"The paper does not mention the number of devices (e.g., GPUs) used for training or inference."
+training_cost,,,not_reported,The paper does not report training cost in terms of monetary value or computational resources.
+training_batch_size,,,not_reported,The paper does not specify the training batch size.
+training_steps_or_epochs,,,not_reported,The paper does not report the number of training steps or epochs.
+precision,,,not_reported,"The paper does not specify the precision (e.g., FP16, FP32) used during training or inference."
+inference_latency,,,not_reported,The paper does not report inference latency.
+inference_throughput,,,not_reported,The paper does not report inference throughput.
+peak_memory,,,not_reported,The paper does not report peak memory usage.
+flops_or_macs,,,not_reported,The paper does not report FLOPs or MACs.
+num_inference_samples,,,not_reported,The paper does not specify the number of inference samples evaluated.
+params,"14m, 91m, 311m",million parameters,stated,"The paper explicitly states the parameter counts for MOIRAI_{Small}, MOIRAI_{Base}, and MOIRAI_{Large} as 14m, 91m, and 311m, respectively."
+context_lengths_evaluated,,,not_reported,"The paper lists multiple context lengths (e.g., 96, 192, 336, 720) in evaluation tables but does not specify a single value for context length."
+horizon_lengths_evaluated,,,not_reported,"The paper lists multiple horizon lengths (e.g., 96, 192, 336, 720) in evaluation tables but does not specify a single value for horizon length."
+inference_batch_size,,,not_reported,The paper does not specify the inference batch size.
diff --git a/result/per_paper/2402.03885/accuracy_efficiency.csv b/result/per_paper/2402.03885/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a27d7de481582d09e2000c16926fa53abab75fe7
--- /dev/null
+++ b/result/per_paper/2402.03885/accuracy_efficiency.csv
@@ -0,0 +1,290 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,GestureMidAirD2,MOMENT_0,MAE,0.608,
+accuracy,UWaveGestureLibraryX,MOMENT_0,MAE,0.821,
+accuracy,GesturePebbleZ2,MOMENT_0,MAE,0.816,
+accuracy,ECG5000,MOMENT_0,MAE,0.942,
+accuracy,OSULeaf,MOMENT_0,MAE,0.785,
+accuracy,MedicalImages,MOMENT_0,MAE,0.762,
+accuracy,Ham,MOMENT_0,MAE,0.581,
+accuracy,DistalPhalanxTW,MOMENT_0,MAE,0.612,
+accuracy,ProximalPhalanxOutlineCorrect,MOMENT_0,MAE,0.856,
+accuracy,FreezerRegularTrain,MOMENT_0,MAE,0.982,
+accuracy,TwoLeadECG,MOMENT_0,MAE,0.847,
+accuracy,GunPointMaleVersusFemale,MOMENT_0,MAE,0.991,
+accuracy,Trace,MOMENT_0,MAE,1.000,
+accuracy,SmoothSubspace,MOMENT_0,MAE,0.820,
+accuracy,MiddlePhalanxTW,MOMENT_0,MAE,0.532,
+accuracy,SyntheticControl,MOMENT_0,MAE,0.990,
+accuracy,ShapesAll,MOMENT_0,MAE,0.815,
+accuracy,AllGestureWiimoteX,MOMENT_0,MAE,0.607,
+accuracy,Wafer,MOMENT_0,MAE,0.997,
+accuracy,FaceFour,MOMENT_0,MAE,0.852,
+accuracy,CricketX,MOMENT_0,MAE,0.749,
+accuracy,DistalPhalanxOutlineCorrect,MOMENT_0,MAE,0.717,
+accuracy,ChlorineConcentration,MOMENT_0,MAE,0.765,
+accuracy,Chinatown,MOMENT_0,MAE,0.965,
+accuracy,GestureMidAirD1,MOMENT_0,MAE,0.646,
+accuracy,MiddlePhalanxOutlineAgeGroup,MOMENT_0,MAE,0.461,
+accuracy,UMD,MOMENT_0,MAE,0.993,
+accuracy,Crop,MOMENT_0,MAE,0.734,
+accuracy,GesturePebbleZ1,MOMENT_0,MAE,0.849,
+accuracy,WordSynonyms,MOMENT_0,MAE,0.688,
+accuracy,ArrowHead,MOMENT_0,MAE,0.743,
+accuracy,Wine,MOMENT_0,MAE,0.537,
+accuracy,Coffee,MOMENT_0,MAE,0.893,
+accuracy,Earthquakes,MOMENT_0,MAE,0.748,
+accuracy,Herring,MOMENT_0,MAE,0.594,
+accuracy,Beef,MOMENT_0,MAE,0.833,
+accuracy,MiddlePhalanxOutlineCorrect,MOMENT_0,MAE,0.467,
+accuracy,ECGFiveDays,MOMENT_0,MAE,0.804,
+accuracy,Yoga,MOMENT_0,MAE,0.834,
+accuracy,Adiac,MOMENT_0,MAE,0.688,
+accuracy,MoteStrain,MOMENT_0,MAE,0.774,
+accuracy,Strawberry,MOMENT_0,MAE,0.951,
+accuracy,InsectWingbeatSound,MOMENT_0,MAE,0.607,
+accuracy,DodgerLoopWeekend,MOMENT_0,MAE,0.826,
+accuracy,Meat,MOMENT_0,MAE,0.917,
+accuracy,MelbournePedestrian,MOMENT_0,MAE,0.876,
+accuracy,FaceAll,MOMENT_0,MAE,0.791,
+accuracy,FacesUCR,MOMENT_0,MAE,0.811,
+accuracy,AllGestureWiimoteY,MOMENT_0,MAE,0.666,
+accuracy,ShakeGestureWiimoteZ,MOMENT_0,MAE,0.960,
+accuracy,BME,MOMENT_0,MAE,0.960,
+accuracy,FordB,MOMENT_0,MAE,0.798,
+accuracy,Fish,MOMENT_0,MAE,0.800,
+accuracy,SonyAIBORobotSurface2,MOMENT_0,MAE,0.829,
+accuracy,FiftyWords,MOMENT_0,MAE,0.802,
+accuracy,ToeSegmentation1,MOMENT_0,MAE,0.925,
+accuracy,FreezerSmallTrain,MOMENT_0,MAE,0.902,
+accuracy,Weather,MOMENT-LP,MSE,0.172,
+accuracy,ETTh1,MOMENT-LP,MSE,0.384,
+accuracy,ETTh2,MOMENT-LP,MSE,0.340,
+accuracy,ETTm1,MOMENT-LP,MSE,0.338,
+accuracy,ETTm2,MOMENT-LP,MSE,0.187,
+accuracy,ILI,MOMENT-LP,MSE,2.317,
+accuracy,ECL,MOMENT-LP,MSE,0.168,
+accuracy,Traffic,MOMENT-LP,MSE,0.593,
+accuracy,Weather,MOMENT_0,MSE,0.105,
+accuracy,Weather,MOMENT_LP,MSE,0.089,
+accuracy,ETTh1,MOMENT_0,MSE,1.008,
+accuracy,ETTh1,MOMENT_LP,MSE,0.602,
+accuracy,ETTh2,MOMENT_0,MSE,0.196,
+accuracy,ETTh2,MOMENT_LP,MSE,0.285,
+accuracy,ETTm1,MOMENT_0,MSE,0.273,
+accuracy,ETTm1,MOMENT_LP,MSE,0.293,
+accuracy,ETTm2,MOMENT_0,MSE,0.087,
+accuracy,ETTm2,MOMENT_LP,MSE,0.164,
+accuracy,Electricity,MOMENT_0,MSE,1.350,
+accuracy,Electricity,MOMENT_LP,MSE,0.818,
+accuracy,lsddb40,MOMENT_0,F_1,0.540,
+accuracy,lsddb40,MOMENT_LP,F_1,0.390,
+accuracy,lsddb40,MOMENT_0,F_1,0.740,
+accuracy,lsddb40,MOMENT_LP,F_1,0.750,
+accuracy,BIDMC1,MOMENT_0,F_1,1.000,
+accuracy,BIDMC1,MOMENT_LP,F_1,1.000,
+accuracy,BIDMC1,MOMENT_0,F_1,0.560,
+accuracy,BIDMC1,MOMENT_LP,F_1,0.650,
+accuracy,CHARISfive,MOMENT_0,F_1,0.130,
+accuracy,CHARISfive,MOMENT_LP,F_1,0.020,
+accuracy,CHARISfive,MOMENT_0,F_1,0.430,
+accuracy,CHARISfive,MOMENT_LP,F_1,0.400,
+accuracy,CHARISTen,MOMENT_0,F_1,0.110,
+accuracy,CHARISTen,MOMENT_LP,F_1,0.040,
+accuracy,CHARISTen,MOMENT_0,F_1,0.500,
+accuracy,CHARISTen,MOMENT_LP,F_1,0.540,
+accuracy,CIMIS44AirTemperature3,MOMENT_0,F_1,0.980,
+accuracy,CIMIS44AirTemperature3,MOMENT_LP,F_1,0.500,
+accuracy,CIMIS44AirTemperature3,MOMENT_0,F_1,0.740,
+accuracy,CIMIS44AirTemperature3,MOMENT_LP,F_1,0.750,
+accuracy,CIMIS44AirTemperature5,MOMENT_0,F_1,0.990,
+accuracy,CIMIS44AirTemperature5,MOMENT_LP,F_1,0.960,
+accuracy,CIMIS44AirTemperature5,MOMENT_0,F_1,0.750,
+accuracy,CIMIS44AirTemperature5,MOMENT_LP,F_1,0.810,
+accuracy,ECG2,MOMENT_0,F_1,1.000,
+accuracy,ECG2,MOMENT_LP,F_1,0.620,
+accuracy,ECG2,MOMENT_0,F_1,0.740,
+accuracy,ECG2,MOMENT_LP,F_1,0.840,
+accuracy,ECG3,MOMENT_0,F_1,0.980,
+accuracy,ECG3,MOMENT_LP,F_1,0.800,
+accuracy,ECG3,MOMENT_0,F_1,0.700,
+accuracy,ECG3,MOMENT_LP,F_1,0.770,
+accuracy,Fantasia,MOMENT_0,F_1,0.950,
+accuracy,Fantasia,MOMENT_LP,F_1,0.660,
+accuracy,Fantasia,MOMENT_0,F_1,0.630,
+accuracy,Fantasia,MOMENT_LP,F_1,0.640,
+accuracy,GP711MarkerLFM5z4,MOMENT_0,F_1,1.000,
+accuracy,GP711MarkerLFM5z4,MOMENT_LP,F_1,0.500,
+accuracy,GP711MarkerLFM5z4,MOMENT_0,F_1,0.630,
+accuracy,GP711MarkerLFM5z4,MOMENT_LP,F_1,0.730,
+accuracy,GP711MarkerLFM5z5,MOMENT_0,F_1,0.970,
+accuracy,GP711MarkerLFM5z5,MOMENT_LP,F_1,0.310,
+accuracy,GP711MarkerLFM5z5,MOMENT_0,F_1,0.760,
+accuracy,GP711MarkerLFM5z5,MOMENT_LP,F_1,0.720,
+accuracy,InternalBleeding4,MOMENT_0,F_1,0.650,
+accuracy,InternalBleeding5,MOMENT_0,F_1,1.000,
+accuracy,InternalBleeding5,MOMENT_LP,F_1,1.000,
+accuracy,InternalBleeding5,MOMENT_0,F_1,0.600,
+accuracy,InternalBleeding5,MOMENT_LP,F_1,0.690,
+accuracy,Italianpowerdemand,MOMENT_0,F_1,0.740,
+accuracy,Italianpowerdemand,MOMENT_LP,F_1,0.590,
+accuracy,Italianpowerdemand,MOMENT_0,F_1,0.800,
+accuracy,Italianpowerdemand,MOMENT_LP,F_1,0.770,
+accuracy,Lab2Cmac011215EPG5,MOMENT_0,F_1,0.980,
+accuracy,Lab2Cmac011215EPG5,MOMENT_LP,F_1,0.340,
+accuracy,Lab2Cmac011215EPG5,MOMENT_0,F_1,0.620,
+accuracy,Lab2Cmac011215EPG5,MOMENT_LP,F_1,0.630,
+accuracy,Lab2Cmac011215EPG6,MOMENT_0,F_1,0.100,
+accuracy,Lab2Cmac011215EPG6,MOMENT_LP,F_1,0.260,
+accuracy,Lab2Cmac011215EPG6,MOMENT_0,F_1,0.480,
+accuracy,Lab2Cmac011215EPG6,MOMENT_LP,F_1,0.480,
+accuracy,MesoplodonDensirostris,MOMENT_0,F_1,0.840,
+accuracy,MesoplodonDensirostris,MOMENT_LP,F_1,0.790,
+accuracy,MesoplodonDensirostris,MOMENT_0,F_1,0.730,
+accuracy,MesoplodonDensirostris,MOMENT_LP,F_1,0.720,
+accuracy,PowerDemand1,MOMENT_0,F_1,0.440,
+accuracy,PowerDemand1,MOMENT_LP,F_1,0.490,
+accuracy,PowerDemand1,MOMENT_0,F_1,0.520,
+accuracy,PowerDemand1,MOMENT_LP,F_1,0.540,
+accuracy,TkeepFirstMARS,MOMENT_0,F_1,0.150,
+accuracy,TkeepFirstMARS,MOMENT_LP,F_1,0.020,
+accuracy,TkeepFirstMARS,MOMENT_0,F_1,0.570,
+accuracy,TkeepFirstMARS,MOMENT_LP,F_1,0.760,
+accuracy,TkeepSecondMARS,MOMENT_0,F_1,1.000,
+accuracy,TkeepSecondMARS,MOMENT_LP,F_1,0.160,
+accuracy,TkeepSecondMARS,MOMENT_0,F_1,0.950,
+accuracy,TkeepSecondMARS,MOMENT_LP,F_1,0.910,
+accuracy,WalkingAceleration5,MOMENT_0,F_1,1.000,
+accuracy,WalkingAceleration5,MOMENT_LP,F_1,0.910,
+accuracy,WalkingAceleration5,MOMENT_0,F_1,0.860,
+accuracy,WalkingAceleration5,MOMENT_LP,F_1,0.870,
+accuracy,apneaecg,MOMENT_0,F_1,0.200,
+accuracy,apneaecg,MOMENT_LP,F_1,0.250,
+accuracy,apneaecg,MOMENT_0,F_1,0.690,
+accuracy,apneaecg,MOMENT_LP,F_1,0.690,
+accuracy,apneaecg2,MOMENT_0,F_1,1.000,
+accuracy,apneaecg2,MOMENT_LP,F_1,1.000,
+accuracy,apneaecg2,MOMENT_0,F_1,0.750,
+accuracy,apneaecg2,MOMENT_LP,F_1,0.740,
+accuracy,gait1,MOMENT_0,F_1,0.360,
+accuracy,gait1,MOMENT_LP,F_1,0.070,
+accuracy,gait1,MOMENT_0,F_1,0.650,
+accuracy,gait1,MOMENT_LP,F_1,0.570,
+accuracy,gaitHunt1,MOMENT_0,F_1,0.430,
+accuracy,gaitHunt1,MOMENT_LP,F_1,0.020,
+accuracy,gaitHunt1,MOMENT_0,F_1,0.640,
+accuracy,gaitHunt1,MOMENT_LP,F_1,0.680,
+accuracy,insectEPG2,MOMENT_0,F_1,0.230,
+accuracy,insectEPG2,MOMENT_LP,F_1,0.140,
+accuracy,insectEPG2,MOMENT_0,F_1,0.570,
+accuracy,insectEPG2,MOMENT_LP,F_1,0.820,
+accuracy,insectEPG4,MOMENT_0,F_1,1.000,
+accuracy,insectEPG4,MOMENT_LP,F_1,0.460,
+accuracy,insectEPG4,MOMENT_0,F_1,0.700,
+accuracy,insectEPG4,MOMENT_LP,F_1,0.720,
+accuracy,ltstdbs30791AS,MOMENT_0,F_1,1.000,
+accuracy,ltstdbs30791AS,MOMENT_LP,F_1,1.000,
+accuracy,ltstdbs30791AS,MOMENT_0,F_1,0.760,
+accuracy,ltstdbs30791AS,MOMENT_LP,F_1,0.810,
+accuracy,mit14046longtermecg,MOMENT_0,F_1,0.590,
+accuracy,mit14046longtermecg,MOMENT_LP,F_1,0.530,
+accuracy,mit14046longtermecg,MOMENT_0,F_1,0.660,
+accuracy,mit14046longtermecg,MOMENT_LP,F_1,0.660,
+accuracy,park3m,MOMENT_0,F_1,0.640,
+accuracy,park3m,MOMENT_LP,F_1,0.200,
+accuracy,park3m,MOMENT_0,F_1,0.750,
+accuracy,park3m,MOMENT_LP,F_1,0.780,
+accuracy,qtdbSel1005V,MOMENT_0,F_1,0.650,
+accuracy,qtdbSel1005V,MOMENT_LP,F_1,0.400,
+accuracy,qtdbSel1005V,MOMENT_0,F_1,0.640,
+accuracy,qtdbSel1005V,MOMENT_LP,F_1,0.640,
+accuracy,qtdbSel100MLII,MOMENT_0,F_1,0.840,
+accuracy,qtdbSel100MLII,MOMENT_LP,F_1,0.410,
+accuracy,qtdbSel100MLII,MOMENT_0,F_1,0.580,
+accuracy,qtdbSel100MLII,MOMENT_LP,F_1,0.620,
+accuracy,resperation1,MOMENT_0,F_1,0.150,
+accuracy,resperation1,MOMENT_LP,F_1,0.030,
+accuracy,resperation1,MOMENT_0,F_1,0.500,
+accuracy,resperation1,MOMENT_LP,F_1,0.670,
+accuracy,s20101mML2,MOMENT_0,F_1,0.710,
+accuracy,s20101mML2,MOMENT_LP,F_1,0.150,
+accuracy,s20101mML2,MOMENT_0,F_1,0.760,
+accuracy,s20101mML2,MOMENT_LP,F_1,0.720,
+accuracy,sddb49,MOMENT_0,F_1,1.000,
+accuracy,sddb49,MOMENT_LP,F_1,0.880,
+accuracy,sddb49,MOMENT_0,F_1,0.730,
+accuracy,sddb49,MOMENT_LP,F_1,0.730,
+accuracy,sel840mECG1,MOMENT_0,F_1,0.660,
+accuracy,sel840mECG1,MOMENT_LP,F_1,0.280,
+accuracy,sel840mECG1,MOMENT_0,F_1,0.720,
+accuracy,sel840mECG1,MOMENT_LP,F_1,0.720,
+accuracy,sel840mECG2,MOMENT_0,F_1,0.390,
+accuracy,sel840mECG2,MOMENT_LP,F_1,0.320,
+accuracy,sel840mECG2,MOMENT_0,F_1,0.710,
+accuracy,sel840mECG2,MOMENT_LP,F_1,0.690,
+accuracy,tilt12744mtable,MOMENT_0,F_1,0.240,
+accuracy,tilt12744mtable,MOMENT_LP,F_1,0.100,
+accuracy,tilt12744mtable,MOMENT_0,F_1,0.670,
+accuracy,tilt12744mtable,MOMENT_LP,F_1,0.740,
+accuracy,tilt12754table,MOMENT_0,F_1,0.640,
+accuracy,tilt12754table,MOMENT_LP,F_1,0.040,
+accuracy,tilt12754table,MOMENT_0,F_1,0.750,
+accuracy,tilt12754table,MOMENT_LP,F_1,0.820,
+accuracy,tiltAPB2,MOMENT_0,F_1,0.980,
+accuracy,tiltAPB2,MOMENT_LP,F_1,0.360,
+accuracy,tiltAPB2,MOMENT_0,F_1,0.750,
+accuracy,tiltAPB2,MOMENT_LP,F_1,0.770,
+accuracy,tiltAPB3,MOMENT_0,F_1,0.850,
+accuracy,tiltAPB3,MOMENT_LP,F_1,0.030,
+accuracy,tiltAPB3,MOMENT_0,F_1,0.610,
+accuracy,tiltAPB3,MOMENT_LP,F_1,0.650,
+accuracy,weallwalk,MOMENT_0,F_1,0.580,
+accuracy,weallwalk,MOMENT_LP,F_1,0.070,
+accuracy,weallwalk,MOMENT_0,F_1,0.930,
+accuracy,weallwalk,MOMENT_LP,F_1,0.930,
+accuracy,ArticularyWordRecognition,MOMENT_0,Accuracy,0.990,
+accuracy,AtrialFibrillation,MOMENT_0,Accuracy,0.200,
+accuracy,BasicMotions,MOMENT_0,Accuracy,1.000,
+accuracy,Cricket,MOMENT_0,Accuracy,0.986,
+accuracy,DuckDuckGeese,MOMENT_0,Accuracy,0.600,
+accuracy,EigenWorms,MOMENT_0,Accuracy,0.809,
+accuracy,Epilepsy,MOMENT_0,Accuracy,0.993,
+accuracy,ERing,MOMENT_0,Accuracy,0.959,
+accuracy,EthanolConcentration,MOMENT_0,Accuracy,0.357,
+accuracy,FaceDetection,MOMENT_0,Accuracy,0.633,
+accuracy,FingerMovements,MOMENT_0,Accuracy,0.490,
+accuracy,HandMovementDirection,MOMENT_0,Accuracy,0.324,
+accuracy,Handwriting,MOMENT_0,Accuracy,0.308,
+accuracy,Heartbeat,MOMENT_0,Accuracy,0.722,
+accuracy,JapaneseVowels,MOMENT_0,Accuracy,0.716,
+accuracy,Libras,MOMENT_0,Accuracy,0.850,
+accuracy,LSST,MOMENT_0,Accuracy,0.411,
+accuracy,MotorImagery,MOMENT_0,Accuracy,0.500,
+accuracy,NATOPS,MOMENT_0,Accuracy,0.828,
+accuracy,PEMS-SF,MOMENT_0,Accuracy,0.896,
+accuracy,PenDigits,MOMENT_0,Accuracy,0.972,
+accuracy,PhonemeSpectra,MOMENT_0,Accuracy,0.233,
+accuracy,RacketSports,MOMENT_0,Accuracy,0.796,
+accuracy,SelfRegulationSCP1,MOMENT_0,Accuracy,0.840,
+accuracy,SelfRegulationSCP2,MOMENT_0,Accuracy,0.478,
+accuracy,SpokenArabicDigits,MOMENT_0,Accuracy,0.981,
+accuracy,StandWalkJump,MOMENT_0,Accuracy,0.400,
+accuracy,UWaveGestureLibrary,MOMENT_0,Accuracy,0.909,
+accuracy,InsectWingbeat,MOMENT_0,Accuracy,0.246,
+accuracy,Mean,MOMENT_0,Accuracy,0.670,
+accuracy,Median,MOMENT_0,Accuracy,0.722,
+accuracy,Std.,MOMENT_0,Accuracy,0.274,
+accuracy,Mean Rank,MOMENT_0,Accuracy,3.466,
+accuracy,Median Rank,MOMENT_0,Accuracy,3.0,
+accuracy,Wins/Losses,MOMENT_0,Accuracy,101.5,
+accuracy,Weather,MOMENT_0,MSE,0.119,
+accuracy,Weather,MOMENT_LP,MSE,0.108,
+accuracy,ETTh1,MOMENT_0,MSE,1.185,
+accuracy,ETTh1,MOMENT_LP,MSE,0.658,
+accuracy,ETTh2,MOMENT_0,MSE,0.225,
+accuracy,ETTh2,MOMENT_LP,MSE,0.304,
+accuracy,ETTm1,MOMENT_0,MSE,0.455,
+accuracy,ETTm1,MOMENT_LP,MSE,0.365,
+accuracy,ETTm2,MOMENT_0,MSE,0.113,
+accuracy,ETTm2,MOMENT_LP,MSE,0.191,
+accuracy,Electricity,MOMENT_0,MSE,1.474,
+accuracy,Electricity,MOMENT_LP,MSE,0.869,
diff --git a/result/per_paper/2402.03885/accuracy_efficiency_traced.csv b/result/per_paper/2402.03885/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..71a048189cd09627a264b8f8ab3be5994b0d07be
--- /dev/null
+++ b/result/per_paper/2402.03885/accuracy_efficiency_traced.csv
@@ -0,0 +1,290 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,GestureMidAirD2,MOMENT_0,MAE,0.608,,0,1,1
+accuracy,UWaveGestureLibraryX,MOMENT_0,MAE,0.821,,0,2,1
+accuracy,GesturePebbleZ2,MOMENT_0,MAE,0.816,,0,3,1
+accuracy,ECG5000,MOMENT_0,MAE,0.942,,0,4,1
+accuracy,OSULeaf,MOMENT_0,MAE,0.785,,0,5,1
+accuracy,MedicalImages,MOMENT_0,MAE,0.762,,0,6,1
+accuracy,Ham,MOMENT_0,MAE,0.581,,0,7,1
+accuracy,DistalPhalanxTW,MOMENT_0,MAE,0.612,,0,8,1
+accuracy,ProximalPhalanxOutlineCorrect,MOMENT_0,MAE,0.856,,0,9,1
+accuracy,FreezerRegularTrain,MOMENT_0,MAE,0.982,,0,10,1
+accuracy,TwoLeadECG,MOMENT_0,MAE,0.847,,0,11,1
+accuracy,GunPointMaleVersusFemale,MOMENT_0,MAE,0.991,,0,12,1
+accuracy,Trace,MOMENT_0,MAE,1.000,,0,13,1
+accuracy,SmoothSubspace,MOMENT_0,MAE,0.820,,0,14,1
+accuracy,MiddlePhalanxTW,MOMENT_0,MAE,0.532,,0,15,1
+accuracy,SyntheticControl,MOMENT_0,MAE,0.990,,0,16,1
+accuracy,ShapesAll,MOMENT_0,MAE,0.815,,0,17,1
+accuracy,AllGestureWiimoteX,MOMENT_0,MAE,0.607,,0,18,1
+accuracy,Wafer,MOMENT_0,MAE,0.997,,0,19,1
+accuracy,FaceFour,MOMENT_0,MAE,0.852,,0,20,1
+accuracy,CricketX,MOMENT_0,MAE,0.749,,0,21,1
+accuracy,DistalPhalanxOutlineCorrect,MOMENT_0,MAE,0.717,,0,22,1
+accuracy,ChlorineConcentration,MOMENT_0,MAE,0.765,,0,23,1
+accuracy,Chinatown,MOMENT_0,MAE,0.965,,0,24,1
+accuracy,GestureMidAirD1,MOMENT_0,MAE,0.646,,0,25,1
+accuracy,MiddlePhalanxOutlineAgeGroup,MOMENT_0,MAE,0.461,,0,26,1
+accuracy,UMD,MOMENT_0,MAE,0.993,,0,27,1
+accuracy,Crop,MOMENT_0,MAE,0.734,,0,28,1
+accuracy,GesturePebbleZ1,MOMENT_0,MAE,0.849,,0,29,1
+accuracy,WordSynonyms,MOMENT_0,MAE,0.688,,0,30,1
+accuracy,ArrowHead,MOMENT_0,MAE,0.743,,0,31,1
+accuracy,Wine,MOMENT_0,MAE,0.537,,0,32,1
+accuracy,Coffee,MOMENT_0,MAE,0.893,,0,33,1
+accuracy,Earthquakes,MOMENT_0,MAE,0.748,,0,34,1
+accuracy,Herring,MOMENT_0,MAE,0.594,,0,35,1
+accuracy,Beef,MOMENT_0,MAE,0.833,,0,36,1
+accuracy,MiddlePhalanxOutlineCorrect,MOMENT_0,MAE,0.467,,0,37,1
+accuracy,ECGFiveDays,MOMENT_0,MAE,0.804,,0,38,1
+accuracy,Yoga,MOMENT_0,MAE,0.834,,0,39,1
+accuracy,Adiac,MOMENT_0,MAE,0.688,,0,40,1
+accuracy,MoteStrain,MOMENT_0,MAE,0.774,,0,41,1
+accuracy,Strawberry,MOMENT_0,MAE,0.951,,0,42,1
+accuracy,InsectWingbeatSound,MOMENT_0,MAE,0.607,,0,43,1
+accuracy,DodgerLoopWeekend,MOMENT_0,MAE,0.826,,0,44,1
+accuracy,Meat,MOMENT_0,MAE,0.917,,0,45,1
+accuracy,MelbournePedestrian,MOMENT_0,MAE,0.876,,0,46,1
+accuracy,FaceAll,MOMENT_0,MAE,0.791,,0,47,1
+accuracy,FacesUCR,MOMENT_0,MAE,0.811,,0,48,1
+accuracy,AllGestureWiimoteY,MOMENT_0,MAE,0.666,,0,49,1
+accuracy,ShakeGestureWiimoteZ,MOMENT_0,MAE,0.960,,0,50,1
+accuracy,BME,MOMENT_0,MAE,0.960,,0,51,1
+accuracy,FordB,MOMENT_0,MAE,0.798,,0,52,1
+accuracy,Fish,MOMENT_0,MAE,0.800,,0,53,1
+accuracy,SonyAIBORobotSurface2,MOMENT_0,MAE,0.829,,0,54,1
+accuracy,FiftyWords,MOMENT_0,MAE,0.802,,0,55,1
+accuracy,ToeSegmentation1,MOMENT_0,MAE,0.925,,0,56,1
+accuracy,FreezerSmallTrain,MOMENT_0,MAE,0.902,,0,57,1
+accuracy,Weather,MOMENT-LP,MSE,0.172,,2,2,12
+accuracy,ETTh1,MOMENT-LP,MSE,0.384,,2,6,12
+accuracy,ETTh2,MOMENT-LP,MSE,0.340,,2,10,12
+accuracy,ETTm1,MOMENT-LP,MSE,0.338,,2,14,12
+accuracy,ETTm2,MOMENT-LP,MSE,0.187,,2,18,12
+accuracy,ILI,MOMENT-LP,MSE,2.317,,2,22,12
+accuracy,ECL,MOMENT-LP,MSE,0.168,,2,26,12
+accuracy,Traffic,MOMENT-LP,MSE,0.593,,2,30,12
+accuracy,Weather,MOMENT_0,MSE,0.105,,4,2,10
+accuracy,Weather,MOMENT_LP,MSE,0.089,,4,2,11
+accuracy,ETTh1,MOMENT_0,MSE,1.008,,4,7,10
+accuracy,ETTh1,MOMENT_LP,MSE,0.602,,4,7,11
+accuracy,ETTh2,MOMENT_0,MSE,0.196,,4,12,10
+accuracy,ETTh2,MOMENT_LP,MSE,0.285,,4,12,11
+accuracy,ETTm1,MOMENT_0,MSE,0.273,,4,17,10
+accuracy,ETTm1,MOMENT_LP,MSE,0.293,,4,17,11
+accuracy,ETTm2,MOMENT_0,MSE,0.087,,4,22,10
+accuracy,ETTm2,MOMENT_LP,MSE,0.164,,4,22,11
+accuracy,Electricity,MOMENT_0,MSE,1.350,,4,27,10
+accuracy,Electricity,MOMENT_LP,MSE,0.818,,4,27,11
+accuracy,lsddb40,MOMENT_0,F_1,0.540,,5,2,3
+accuracy,lsddb40,MOMENT_LP,F_1,0.390,,5,2,4
+accuracy,lsddb40,MOMENT_0,F_1,0.740,,5,2,8
+accuracy,lsddb40,MOMENT_LP,F_1,0.750,,5,2,9
+accuracy,BIDMC1,MOMENT_0,F_1,1.000,,5,3,3
+accuracy,BIDMC1,MOMENT_LP,F_1,1.000,,5,3,4
+accuracy,BIDMC1,MOMENT_0,F_1,0.560,,5,3,8
+accuracy,BIDMC1,MOMENT_LP,F_1,0.650,,5,3,9
+accuracy,CHARISfive,MOMENT_0,F_1,0.130,,5,4,3
+accuracy,CHARISfive,MOMENT_LP,F_1,0.020,,5,4,4
+accuracy,CHARISfive,MOMENT_0,F_1,0.430,,5,4,8
+accuracy,CHARISfive,MOMENT_LP,F_1,0.400,,5,4,9
+accuracy,CHARISTen,MOMENT_0,F_1,0.110,,5,5,3
+accuracy,CHARISTen,MOMENT_LP,F_1,0.040,,5,5,4
+accuracy,CHARISTen,MOMENT_0,F_1,0.500,,5,5,8
+accuracy,CHARISTen,MOMENT_LP,F_1,0.540,,5,5,9
+accuracy,CIMIS44AirTemperature3,MOMENT_0,F_1,0.980,,5,6,3
+accuracy,CIMIS44AirTemperature3,MOMENT_LP,F_1,0.500,,5,6,4
+accuracy,CIMIS44AirTemperature3,MOMENT_0,F_1,0.740,,5,6,8
+accuracy,CIMIS44AirTemperature3,MOMENT_LP,F_1,0.750,,5,6,9
+accuracy,CIMIS44AirTemperature5,MOMENT_0,F_1,0.990,,5,7,3
+accuracy,CIMIS44AirTemperature5,MOMENT_LP,F_1,0.960,,5,7,4
+accuracy,CIMIS44AirTemperature5,MOMENT_0,F_1,0.750,,5,7,8
+accuracy,CIMIS44AirTemperature5,MOMENT_LP,F_1,0.810,,5,7,9
+accuracy,ECG2,MOMENT_0,F_1,1.000,,5,8,3
+accuracy,ECG2,MOMENT_LP,F_1,0.620,,5,8,4
+accuracy,ECG2,MOMENT_0,F_1,0.740,,5,8,8
+accuracy,ECG2,MOMENT_LP,F_1,0.840,,5,8,9
+accuracy,ECG3,MOMENT_0,F_1,0.980,,5,9,3
+accuracy,ECG3,MOMENT_LP,F_1,0.800,,5,9,4
+accuracy,ECG3,MOMENT_0,F_1,0.700,,5,9,8
+accuracy,ECG3,MOMENT_LP,F_1,0.770,,5,9,9
+accuracy,Fantasia,MOMENT_0,F_1,0.950,,5,10,3
+accuracy,Fantasia,MOMENT_LP,F_1,0.660,,5,10,4
+accuracy,Fantasia,MOMENT_0,F_1,0.630,,5,10,8
+accuracy,Fantasia,MOMENT_LP,F_1,0.640,,5,10,9
+accuracy,GP711MarkerLFM5z4,MOMENT_0,F_1,1.000,,5,11,3
+accuracy,GP711MarkerLFM5z4,MOMENT_LP,F_1,0.500,,5,11,4
+accuracy,GP711MarkerLFM5z4,MOMENT_0,F_1,0.630,,5,11,8
+accuracy,GP711MarkerLFM5z4,MOMENT_LP,F_1,0.730,,5,11,9
+accuracy,GP711MarkerLFM5z5,MOMENT_0,F_1,0.970,,5,12,3
+accuracy,GP711MarkerLFM5z5,MOMENT_LP,F_1,0.310,,5,12,4
+accuracy,GP711MarkerLFM5z5,MOMENT_0,F_1,0.760,,5,12,8
+accuracy,GP711MarkerLFM5z5,MOMENT_LP,F_1,0.720,,5,12,9
+accuracy,InternalBleeding4,MOMENT_0,F_1,0.650,,5,13,8
+accuracy,InternalBleeding5,MOMENT_0,F_1,1.000,,5,14,3
+accuracy,InternalBleeding5,MOMENT_LP,F_1,1.000,,5,14,4
+accuracy,InternalBleeding5,MOMENT_0,F_1,0.600,,5,14,8
+accuracy,InternalBleeding5,MOMENT_LP,F_1,0.690,,5,14,9
+accuracy,Italianpowerdemand,MOMENT_0,F_1,0.740,,5,15,3
+accuracy,Italianpowerdemand,MOMENT_LP,F_1,0.590,,5,15,4
+accuracy,Italianpowerdemand,MOMENT_0,F_1,0.800,,5,15,8
+accuracy,Italianpowerdemand,MOMENT_LP,F_1,0.770,,5,15,9
+accuracy,Lab2Cmac011215EPG5,MOMENT_0,F_1,0.980,,5,16,3
+accuracy,Lab2Cmac011215EPG5,MOMENT_LP,F_1,0.340,,5,16,4
+accuracy,Lab2Cmac011215EPG5,MOMENT_0,F_1,0.620,,5,16,8
+accuracy,Lab2Cmac011215EPG5,MOMENT_LP,F_1,0.630,,5,16,9
+accuracy,Lab2Cmac011215EPG6,MOMENT_0,F_1,0.100,,5,17,3
+accuracy,Lab2Cmac011215EPG6,MOMENT_LP,F_1,0.260,,5,17,4
+accuracy,Lab2Cmac011215EPG6,MOMENT_0,F_1,0.480,,5,17,8
+accuracy,Lab2Cmac011215EPG6,MOMENT_LP,F_1,0.480,,5,17,9
+accuracy,MesoplodonDensirostris,MOMENT_0,F_1,0.840,,5,18,3
+accuracy,MesoplodonDensirostris,MOMENT_LP,F_1,0.790,,5,18,4
+accuracy,MesoplodonDensirostris,MOMENT_0,F_1,0.730,,5,18,8
+accuracy,MesoplodonDensirostris,MOMENT_LP,F_1,0.720,,5,18,9
+accuracy,PowerDemand1,MOMENT_0,F_1,0.440,,5,19,3
+accuracy,PowerDemand1,MOMENT_LP,F_1,0.490,,5,19,4
+accuracy,PowerDemand1,MOMENT_0,F_1,0.520,,5,19,8
+accuracy,PowerDemand1,MOMENT_LP,F_1,0.540,,5,19,9
+accuracy,TkeepFirstMARS,MOMENT_0,F_1,0.150,,5,20,3
+accuracy,TkeepFirstMARS,MOMENT_LP,F_1,0.020,,5,20,4
+accuracy,TkeepFirstMARS,MOMENT_0,F_1,0.570,,5,20,8
+accuracy,TkeepFirstMARS,MOMENT_LP,F_1,0.760,,5,20,9
+accuracy,TkeepSecondMARS,MOMENT_0,F_1,1.000,,5,21,3
+accuracy,TkeepSecondMARS,MOMENT_LP,F_1,0.160,,5,21,4
+accuracy,TkeepSecondMARS,MOMENT_0,F_1,0.950,,5,21,8
+accuracy,TkeepSecondMARS,MOMENT_LP,F_1,0.910,,5,21,9
+accuracy,WalkingAceleration5,MOMENT_0,F_1,1.000,,5,22,3
+accuracy,WalkingAceleration5,MOMENT_LP,F_1,0.910,,5,22,4
+accuracy,WalkingAceleration5,MOMENT_0,F_1,0.860,,5,22,8
+accuracy,WalkingAceleration5,MOMENT_LP,F_1,0.870,,5,22,9
+accuracy,apneaecg,MOMENT_0,F_1,0.200,,5,23,3
+accuracy,apneaecg,MOMENT_LP,F_1,0.250,,5,23,4
+accuracy,apneaecg,MOMENT_0,F_1,0.690,,5,23,8
+accuracy,apneaecg,MOMENT_LP,F_1,0.690,,5,23,9
+accuracy,apneaecg2,MOMENT_0,F_1,1.000,,5,24,3
+accuracy,apneaecg2,MOMENT_LP,F_1,1.000,,5,24,4
+accuracy,apneaecg2,MOMENT_0,F_1,0.750,,5,24,8
+accuracy,apneaecg2,MOMENT_LP,F_1,0.740,,5,24,9
+accuracy,gait1,MOMENT_0,F_1,0.360,,5,25,3
+accuracy,gait1,MOMENT_LP,F_1,0.070,,5,25,4
+accuracy,gait1,MOMENT_0,F_1,0.650,,5,25,8
+accuracy,gait1,MOMENT_LP,F_1,0.570,,5,25,9
+accuracy,gaitHunt1,MOMENT_0,F_1,0.430,,5,26,3
+accuracy,gaitHunt1,MOMENT_LP,F_1,0.020,,5,26,4
+accuracy,gaitHunt1,MOMENT_0,F_1,0.640,,5,26,8
+accuracy,gaitHunt1,MOMENT_LP,F_1,0.680,,5,26,9
+accuracy,insectEPG2,MOMENT_0,F_1,0.230,,5,27,3
+accuracy,insectEPG2,MOMENT_LP,F_1,0.140,,5,27,4
+accuracy,insectEPG2,MOMENT_0,F_1,0.570,,5,27,8
+accuracy,insectEPG2,MOMENT_LP,F_1,0.820,,5,27,9
+accuracy,insectEPG4,MOMENT_0,F_1,1.000,,5,28,3
+accuracy,insectEPG4,MOMENT_LP,F_1,0.460,,5,28,4
+accuracy,insectEPG4,MOMENT_0,F_1,0.700,,5,28,8
+accuracy,insectEPG4,MOMENT_LP,F_1,0.720,,5,28,9
+accuracy,ltstdbs30791AS,MOMENT_0,F_1,1.000,,5,29,3
+accuracy,ltstdbs30791AS,MOMENT_LP,F_1,1.000,,5,29,4
+accuracy,ltstdbs30791AS,MOMENT_0,F_1,0.760,,5,29,8
+accuracy,ltstdbs30791AS,MOMENT_LP,F_1,0.810,,5,29,9
+accuracy,mit14046longtermecg,MOMENT_0,F_1,0.590,,5,30,3
+accuracy,mit14046longtermecg,MOMENT_LP,F_1,0.530,,5,30,4
+accuracy,mit14046longtermecg,MOMENT_0,F_1,0.660,,5,30,8
+accuracy,mit14046longtermecg,MOMENT_LP,F_1,0.660,,5,30,9
+accuracy,park3m,MOMENT_0,F_1,0.640,,5,31,3
+accuracy,park3m,MOMENT_LP,F_1,0.200,,5,31,4
+accuracy,park3m,MOMENT_0,F_1,0.750,,5,31,8
+accuracy,park3m,MOMENT_LP,F_1,0.780,,5,31,9
+accuracy,qtdbSel1005V,MOMENT_0,F_1,0.650,,5,32,3
+accuracy,qtdbSel1005V,MOMENT_LP,F_1,0.400,,5,32,4
+accuracy,qtdbSel1005V,MOMENT_0,F_1,0.640,,5,32,8
+accuracy,qtdbSel1005V,MOMENT_LP,F_1,0.640,,5,32,9
+accuracy,qtdbSel100MLII,MOMENT_0,F_1,0.840,,5,33,3
+accuracy,qtdbSel100MLII,MOMENT_LP,F_1,0.410,,5,33,4
+accuracy,qtdbSel100MLII,MOMENT_0,F_1,0.580,,5,33,8
+accuracy,qtdbSel100MLII,MOMENT_LP,F_1,0.620,,5,33,9
+accuracy,resperation1,MOMENT_0,F_1,0.150,,5,34,3
+accuracy,resperation1,MOMENT_LP,F_1,0.030,,5,34,4
+accuracy,resperation1,MOMENT_0,F_1,0.500,,5,34,8
+accuracy,resperation1,MOMENT_LP,F_1,0.670,,5,34,9
+accuracy,s20101mML2,MOMENT_0,F_1,0.710,,5,35,3
+accuracy,s20101mML2,MOMENT_LP,F_1,0.150,,5,35,4
+accuracy,s20101mML2,MOMENT_0,F_1,0.760,,5,35,8
+accuracy,s20101mML2,MOMENT_LP,F_1,0.720,,5,35,9
+accuracy,sddb49,MOMENT_0,F_1,1.000,,5,36,3
+accuracy,sddb49,MOMENT_LP,F_1,0.880,,5,36,4
+accuracy,sddb49,MOMENT_0,F_1,0.730,,5,36,8
+accuracy,sddb49,MOMENT_LP,F_1,0.730,,5,36,9
+accuracy,sel840mECG1,MOMENT_0,F_1,0.660,,5,37,3
+accuracy,sel840mECG1,MOMENT_LP,F_1,0.280,,5,37,4
+accuracy,sel840mECG1,MOMENT_0,F_1,0.720,,5,37,8
+accuracy,sel840mECG1,MOMENT_LP,F_1,0.720,,5,37,9
+accuracy,sel840mECG2,MOMENT_0,F_1,0.390,,5,38,3
+accuracy,sel840mECG2,MOMENT_LP,F_1,0.320,,5,38,4
+accuracy,sel840mECG2,MOMENT_0,F_1,0.710,,5,38,8
+accuracy,sel840mECG2,MOMENT_LP,F_1,0.690,,5,38,9
+accuracy,tilt12744mtable,MOMENT_0,F_1,0.240,,5,39,3
+accuracy,tilt12744mtable,MOMENT_LP,F_1,0.100,,5,39,4
+accuracy,tilt12744mtable,MOMENT_0,F_1,0.670,,5,39,8
+accuracy,tilt12744mtable,MOMENT_LP,F_1,0.740,,5,39,9
+accuracy,tilt12754table,MOMENT_0,F_1,0.640,,5,40,3
+accuracy,tilt12754table,MOMENT_LP,F_1,0.040,,5,40,4
+accuracy,tilt12754table,MOMENT_0,F_1,0.750,,5,40,8
+accuracy,tilt12754table,MOMENT_LP,F_1,0.820,,5,40,9
+accuracy,tiltAPB2,MOMENT_0,F_1,0.980,,5,41,3
+accuracy,tiltAPB2,MOMENT_LP,F_1,0.360,,5,41,4
+accuracy,tiltAPB2,MOMENT_0,F_1,0.750,,5,41,8
+accuracy,tiltAPB2,MOMENT_LP,F_1,0.770,,5,41,9
+accuracy,tiltAPB3,MOMENT_0,F_1,0.850,,5,42,3
+accuracy,tiltAPB3,MOMENT_LP,F_1,0.030,,5,42,4
+accuracy,tiltAPB3,MOMENT_0,F_1,0.610,,5,42,8
+accuracy,tiltAPB3,MOMENT_LP,F_1,0.650,,5,42,9
+accuracy,weallwalk,MOMENT_0,F_1,0.580,,5,43,3
+accuracy,weallwalk,MOMENT_LP,F_1,0.070,,5,43,4
+accuracy,weallwalk,MOMENT_0,F_1,0.930,,5,43,8
+accuracy,weallwalk,MOMENT_LP,F_1,0.930,,5,43,9
+accuracy,ArticularyWordRecognition,MOMENT_0,Accuracy,0.990,,6,1,1
+accuracy,AtrialFibrillation,MOMENT_0,Accuracy,0.200,,6,2,1
+accuracy,BasicMotions,MOMENT_0,Accuracy,1.000,,6,3,1
+accuracy,Cricket,MOMENT_0,Accuracy,0.986,,6,4,1
+accuracy,DuckDuckGeese,MOMENT_0,Accuracy,0.600,,6,5,1
+accuracy,EigenWorms,MOMENT_0,Accuracy,0.809,,6,6,1
+accuracy,Epilepsy,MOMENT_0,Accuracy,0.993,,6,7,1
+accuracy,ERing,MOMENT_0,Accuracy,0.959,,6,8,1
+accuracy,EthanolConcentration,MOMENT_0,Accuracy,0.357,,6,9,1
+accuracy,FaceDetection,MOMENT_0,Accuracy,0.633,,6,10,1
+accuracy,FingerMovements,MOMENT_0,Accuracy,0.490,,6,11,1
+accuracy,HandMovementDirection,MOMENT_0,Accuracy,0.324,,6,12,1
+accuracy,Handwriting,MOMENT_0,Accuracy,0.308,,6,13,1
+accuracy,Heartbeat,MOMENT_0,Accuracy,0.722,,6,14,1
+accuracy,JapaneseVowels,MOMENT_0,Accuracy,0.716,,6,15,1
+accuracy,Libras,MOMENT_0,Accuracy,0.850,,6,16,1
+accuracy,LSST,MOMENT_0,Accuracy,0.411,,6,17,1
+accuracy,MotorImagery,MOMENT_0,Accuracy,0.500,,6,18,1
+accuracy,NATOPS,MOMENT_0,Accuracy,0.828,,6,19,1
+accuracy,PEMS-SF,MOMENT_0,Accuracy,0.896,,6,20,1
+accuracy,PenDigits,MOMENT_0,Accuracy,0.972,,6,21,1
+accuracy,PhonemeSpectra,MOMENT_0,Accuracy,0.233,,6,22,1
+accuracy,RacketSports,MOMENT_0,Accuracy,0.796,,6,23,1
+accuracy,SelfRegulationSCP1,MOMENT_0,Accuracy,0.840,,6,24,1
+accuracy,SelfRegulationSCP2,MOMENT_0,Accuracy,0.478,,6,25,1
+accuracy,SpokenArabicDigits,MOMENT_0,Accuracy,0.981,,6,26,1
+accuracy,StandWalkJump,MOMENT_0,Accuracy,0.400,,6,27,1
+accuracy,UWaveGestureLibrary,MOMENT_0,Accuracy,0.909,,6,28,1
+accuracy,InsectWingbeat,MOMENT_0,Accuracy,0.246,,6,29,1
+accuracy,Mean,MOMENT_0,Accuracy,0.670,,6,30,1
+accuracy,Median,MOMENT_0,Accuracy,0.722,,6,31,1
+accuracy,Std.,MOMENT_0,Accuracy,0.274,,6,32,1
+accuracy,Mean Rank,MOMENT_0,Accuracy,3.466,,6,33,1
+accuracy,Median Rank,MOMENT_0,Accuracy,3.0,,6,34,1
+accuracy,Wins/Losses,MOMENT_0,Accuracy,101.5,,6,35,1
+accuracy,Weather,MOMENT_0,MSE,0.119,,7,2,9
+accuracy,Weather,MOMENT_LP,MSE,0.108,,7,2,10
+accuracy,ETTh1,MOMENT_0,MSE,1.185,,7,3,9
+accuracy,ETTh1,MOMENT_LP,MSE,0.658,,7,3,10
+accuracy,ETTh2,MOMENT_0,MSE,0.225,,7,4,9
+accuracy,ETTh2,MOMENT_LP,MSE,0.304,,7,4,10
+accuracy,ETTm1,MOMENT_0,MSE,0.455,,7,5,9
+accuracy,ETTm1,MOMENT_LP,MSE,0.365,,7,5,10
+accuracy,ETTm2,MOMENT_0,MSE,0.113,,7,6,9
+accuracy,ETTm2,MOMENT_LP,MSE,0.191,,7,6,10
+accuracy,Electricity,MOMENT_0,MSE,1.474,,7,7,9
+accuracy,Electricity,MOMENT_LP,MSE,0.869,,7,7,10
diff --git a/result/per_paper/2402.03885/components_architecture.csv b/result/per_paper/2402.03885/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ea1ed18d52f9a66d0d769507f8b203d39e71c04b
--- /dev/null
+++ b/result/per_paper/2402.03885/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Time Series Pile,A large and diverse collection of public time series data compiled to enable multi-dataset pretraining for time series foundation models.,proposed_here,,"we compile a large and diverse collection of public time series, called the Time series Pile"
+Transformer with Patching,A modified transformer architecture that processes time series data by treating sub-sequences (patches) as input tokens instead of individual time points.,proposed_here,"Nie et al., 2023",we build on top of the transformer architecture which takes disjoint time series sub-sequences (or patches) as input
+Masked Prediction for Pretraining,"A pretraining task where time series patches are masked using a special embedding, and the model learns to reconstruct them, enabling effective representation learning for forecasting and imputation.",reused_cited,,"we use the masked prediction task to pretrain our model, using a special embedding [...] to mask time series patches instead of zeros"
+Cross-Modal Transfer Learning,"A technique leveraging pre-trained language models (LLMs) to transfer knowledge to time series tasks, enabling sequence modeling capabilities in the time series domain.",reused_cited,"Lu et al., 2022",Lu et al. (2022) had first shown that transformers pre-trained on text data (LLMs) can effectively solve sequence modeling tasks in
diff --git a/result/per_paper/2402.03885/computational.csv b/result/per_paper/2402.03885/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..cb65ad264914c5d2365da681ec9dbae69b171493
--- /dev/null
+++ b/result/per_paper/2402.03885/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No explicit mention of hardware type (e.g., GPU/TPU)."
+num_devices,,,not_reported,No explicit mention of number of devices used.
+training_cost,,,not_reported,"No explicit mention of training cost (e.g., FLOPs, energy, or monetary cost)."
+training_batch_size,2048,batch,stated,Trained with a batch size of 2048.
+training_steps_or_epochs,,,not_reported,No explicit mention of training steps or epochs.
+precision,bfloat-16 (mixed precision),,stated,Trained in mixed precision using float-32 for unstable operations and bfloat-16 otherwise.
+inference_latency,,,not_reported,No explicit mention of inference latency.
+inference_throughput,,,not_reported,No explicit mention of inference throughput.
+peak_memory,,,not_reported,No explicit mention of peak memory usage.
+flops_or_macs,,,not_reported,No explicit mention of FLOPs or MACs.
+num_inference_samples,,,not_reported,No explicit mention of inference sample count.
+params,,,not_reported,No explicit mention of parameter count.
+context_lengths_evaluated,,,not_reported,No explicit mention of model-specific context lengths evaluated.
+horizon_lengths_evaluated,,,not_reported,"Task-specific forecast horizons (e.g., H = {96, 720}) are described, but not the model's own horizon."
+inference_batch_size,,,not_reported,No explicit mention of inference batch size.
diff --git a/result/per_paper/2402.07570/accuracy_efficiency.csv b/result/per_paper/2402.07570/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..8f4966d2c6368e2a6719f28aa1f44e9bbbe5f070
--- /dev/null
+++ b/result/per_paper/2402.07570/accuracy_efficiency.csv
@@ -0,0 +1,54 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTm1,GTT-Large,MSE,0.171,
+accuracy,mean,GTT-Large,MSE,0.322,
+accuracy,ETTm2,GTT-Large,MSE,0.565,
+accuracy,mean,GTT-Large,MSE,0.598,
+accuracy,ETTh1,GTT-Large,MSE,0.198,
+accuracy,mean,GTT-Large,MSE,0.350,
+accuracy,ETTh2,GTT-Large,MSE,0.608,
+accuracy,mean,GTT-Large,MSE,0.618,
+accuracy,Electricity,GTT-Large,MSE,2.206,
+accuracy,mean,GTT-Large,MSE,1.233,
+accuracy,Traffic,GTT-Large,MSE,4.173,
+accuracy,mean,GTT-Large,MSE,1.650,
+accuracy,Weather,GTT-Large,MSE,0.002,
+accuracy,mean,GTT-Large,MSE,0.040,
+accuracy,ILI,GTT-Large,MSE,1.102,
+accuracy,mean,GTT-Large,MSE,0.940,
+accuracy,ETTm1,GTT (ZS),MSE,0.448,
+accuracy,ETTm1,GTT (FT),MSE,0.452,
+accuracy,ETTm2,GTT (ZS),MSE,0.305,
+accuracy,ETTm2,GTT (FT),MSE,0.349,
+accuracy,ETTh1,GTT (ZS),MSE,0.440,
+accuracy,ETTh1,GTT (FT),MSE,0.460,
+accuracy,ETTh2,GTT (ZS),MSE,0.434,
+accuracy,ETTh2,GTT (FT),MSE,0.447,
+accuracy,Electricity,GTT (ZS),MSE,0.214,
+accuracy,Electricity,GTT (FT),MSE,0.327,
+accuracy,Traffic,GTT (ZS),MSE,0.610,
+accuracy,Traffic,GTT (FT),MSE,0.376,
+accuracy,Weather,GTT (ZS),MSE,0.309,
+accuracy,Weather,GTT (FT),MSE,0.360,
+accuracy,ILI,GTT (ZS),MSE,2.847,
+accuracy,ILI,GTT (FT),MSE,1.144,
+accuracy,ETTm1,GTT (General Time Transformer),MSE,0.160,
+accuracy,ETTm1,ForecastPFN,MSE,0.175,
+accuracy,ETTm2,GTT (General Time Transformer),MSE,0.254,
+accuracy,ETTm2,ForecastPFN,MSE,0.568,
+accuracy,ETTh1,GTT (General Time Transformer),MSE,0.222,
+accuracy,ETTh1,ForecastPFN,MSE,0.190,
+accuracy,ETTh2,GTT (General Time Transformer),MSE,0.359,
+accuracy,ETTh2,ForecastPFN,MSE,0.604,
+accuracy,Electricity,GTT (General Time Transformer),MSE,0.355,
+accuracy,Electricity,ForecastPFN,MSE,2.257,
+accuracy,Traffic,GTT (General Time Transformer),MSE,0.219,
+accuracy,Traffic,ForecastPFN,MSE,4.121,
+accuracy,Weather,GTT (General Time Transformer),MSE,0.028,
+accuracy,Weather,ForecastPFN,MSE,0.003,
+accuracy,ILI,GTT (General Time Transformer),MSE,0.551,
+accuracy,ILI,ForecastPFN,MSE,2.903,
+accuracy,ETTm1,GTT (General Time Transformer),NRMSE,0.591,
+accuracy,ETTm2,GTT (General Time Transformer),NRMSE,0.199,
+accuracy,ETTh1,GTT (General Time Transformer),NRMSE,0.672,
+accuracy,ETTh2,GTT (General Time Transformer),NRMSE,0.238,
+accuracy,ILI,GTT (General Time Transformer),NRMSE,0.477,
diff --git a/result/per_paper/2402.07570/accuracy_efficiency_traced.csv b/result/per_paper/2402.07570/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a4e284ebb5c9cb2cf0f125fc2546c1d327db3093
--- /dev/null
+++ b/result/per_paper/2402.07570/accuracy_efficiency_traced.csv
@@ -0,0 +1,54 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTm1,GTT-Large,MSE,0.171,,1,2,6
+accuracy,mean,GTT-Large,MSE,0.322,,1,6,6
+accuracy,ETTm2,GTT-Large,MSE,0.565,,1,7,6
+accuracy,mean,GTT-Large,MSE,0.598,,1,11,6
+accuracy,ETTh1,GTT-Large,MSE,0.198,,1,12,6
+accuracy,mean,GTT-Large,MSE,0.350,,1,16,6
+accuracy,ETTh2,GTT-Large,MSE,0.608,,1,17,6
+accuracy,mean,GTT-Large,MSE,0.618,,1,21,6
+accuracy,Electricity,GTT-Large,MSE,2.206,,1,22,6
+accuracy,mean,GTT-Large,MSE,1.233,,1,26,6
+accuracy,Traffic,GTT-Large,MSE,4.173,,1,27,6
+accuracy,mean,GTT-Large,MSE,1.650,,1,31,6
+accuracy,Weather,GTT-Large,MSE,0.002,,1,32,6
+accuracy,mean,GTT-Large,MSE,0.040,,1,36,6
+accuracy,ILI,GTT-Large,MSE,1.102,,1,37,6
+accuracy,mean,GTT-Large,MSE,0.940,,1,41,6
+accuracy,ETTm1,GTT (ZS),MSE,0.448,,2,2,11
+accuracy,ETTm1,GTT (FT),MSE,0.452,,2,2,12
+accuracy,ETTm2,GTT (ZS),MSE,0.305,,2,3,11
+accuracy,ETTm2,GTT (FT),MSE,0.349,,2,3,12
+accuracy,ETTh1,GTT (ZS),MSE,0.440,,2,4,11
+accuracy,ETTh1,GTT (FT),MSE,0.460,,2,4,12
+accuracy,ETTh2,GTT (ZS),MSE,0.434,,2,5,11
+accuracy,ETTh2,GTT (FT),MSE,0.447,,2,5,12
+accuracy,Electricity,GTT (ZS),MSE,0.214,,2,6,11
+accuracy,Electricity,GTT (FT),MSE,0.327,,2,6,12
+accuracy,Traffic,GTT (ZS),MSE,0.610,,2,7,11
+accuracy,Traffic,GTT (FT),MSE,0.376,,2,7,12
+accuracy,Weather,GTT (ZS),MSE,0.309,,2,8,11
+accuracy,Weather,GTT (FT),MSE,0.360,,2,8,12
+accuracy,ILI,GTT (ZS),MSE,2.847,,2,9,11
+accuracy,ILI,GTT (FT),MSE,1.144,,2,9,12
+accuracy,ETTm1,GTT (General Time Transformer),MSE,0.160,,3,2,2
+accuracy,ETTm1,ForecastPFN,MSE,0.175,,3,2,3
+accuracy,ETTm2,GTT (General Time Transformer),MSE,0.254,,3,3,2
+accuracy,ETTm2,ForecastPFN,MSE,0.568,,3,3,3
+accuracy,ETTh1,GTT (General Time Transformer),MSE,0.222,,3,4,2
+accuracy,ETTh1,ForecastPFN,MSE,0.190,,3,4,3
+accuracy,ETTh2,GTT (General Time Transformer),MSE,0.359,,3,5,2
+accuracy,ETTh2,ForecastPFN,MSE,0.604,,3,5,3
+accuracy,Electricity,GTT (General Time Transformer),MSE,0.355,,3,6,2
+accuracy,Electricity,ForecastPFN,MSE,2.257,,3,6,3
+accuracy,Traffic,GTT (General Time Transformer),MSE,0.219,,3,7,2
+accuracy,Traffic,ForecastPFN,MSE,4.121,,3,7,3
+accuracy,Weather,GTT (General Time Transformer),MSE,0.028,,3,8,2
+accuracy,Weather,ForecastPFN,MSE,0.003,,3,8,3
+accuracy,ILI,GTT (General Time Transformer),MSE,0.551,,3,9,2
+accuracy,ILI,ForecastPFN,MSE,2.903,,3,9,3
+accuracy,ETTm1,GTT (General Time Transformer),NRMSE,0.591,,4,2,3
+accuracy,ETTm2,GTT (General Time Transformer),NRMSE,0.199,,4,3,3
+accuracy,ETTh1,GTT (General Time Transformer),NRMSE,0.672,,4,4,3
+accuracy,ETTh2,GTT (General Time Transformer),NRMSE,0.238,,4,5,3
+accuracy,ILI,GTT (General Time Transformer),NRMSE,0.477,,4,6,3
diff --git a/result/per_paper/2402.07570/components_architecture.csv b/result/per_paper/2402.07570/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..16710121e68249535a469d7fedd0f63495c4853f
--- /dev/null
+++ b/result/per_paper/2402.07570/components_architecture.csv
@@ -0,0 +1,8 @@
+component,what_it_is,provenance,citation,evidence
+Encoder-Only Architecture,"A model structure that uses only the encoder component of the Transformer, without a decoder, for processing time series data.",proposed_here,,We adopt an encoder-only architecture for GTT with the fewest possible modifications to the standard Transformer.
+Cross-Channel Attention Stage,A mechanism added after the temporal attention stage in each multi-head self-attention block to capture dependencies between different channels (variables) in the time series.,proposed_here,,The only major modification we introduced is a cross-channel attention stage after the temporal attention stage in each multi-head self-attention block to capture cross-variate dependency between channels.
+Temporal Attention Stage,A component within the multi-head self-attention block that focuses on temporal relationships within a single channel of the time series.,reused_cited,"Vaswani et al., 2017",The only major modification we introduced is a cross-channel attention stage after the temporal attention stage in each multi-head self-attention block...
+Multi-Head Self-Attention Blocks,"A standard Transformer component that allows the model to attend to information across different positions in the input sequence, with multiple attention heads to capture diverse patterns.",reused_cited,"Vaswani et al., 2017",We adopt an encoder-only architecture for GTT with the fewest possible modifications to the standard Transformer.
+Auto-Regressive Approach,"A method for generating predictions sequentially, where each prediction depends on previous outputs, used to handle long-term forecasting beyond M time steps.",proposed_here,,GTT employs an auto-regressive approach to handle long-term forecasting tasks extend beyond M time steps.
+Curve Shape Representation,"A representation of time series data as sequences of non-overlapping curve shapes, each with a unified numerical magnitude, to abstract away raw values and focus on shape patterns.",proposed_here,,each time series sample is represented as a sequence of non-overlapping curve shapes with a unified numerical magnitude.
+Context Window of N Preceding Curve Shapes,A mechanism where the model uses N preceding curve shapes as input context to predict the next curve shape in a channel-wise manner.,proposed_here,,GTT is trained to use N preceding curve shapes as the context to predict the next curve shape on a channel-wise basis.
diff --git a/result/per_paper/2402.07570/computational.csv b/result/per_paper/2402.07570/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..33b67aa6cd503179060213fa9e9c68b1dda051d5
--- /dev/null
+++ b/result/per_paper/2402.07570/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No specific hardware (e.g., GPU/TPU) is mentioned."
+num_devices,,,not_reported,No information on the number of devices used for training or inference.
+training_cost,,,not_reported,"No monetary or resource cost (e.g., cloud credits, energy) is reported."
+training_batch_size,,,not_reported,Batch size is not explicitly stated.
+training_steps_or_epochs,,,not_reported,"Training stops when validation loss increases for three consecutive epochs, but total steps/epochs are not quantified."
+precision,,,not_reported,"No mention of training/inference precision (e.g., FP16, FP32)."
+inference_latency,,,not_reported,"No latency measurements (e.g., ms per sample) are provided."
+inference_throughput,,,not_reported,"No throughput (e.g., samples/sec) is reported."
+peak_memory,,,not_reported,"No peak memory usage (e.g., GB) is mentioned."
+flops_or_macs,,,not_reported,No FLOPs or MACs are quantified.
+num_inference_samples,,,not_reported,No explicit count of inference samples is provided.
+params,57000000,parameters,stated,"The largest model has 57M parameters (""Our largest trained model has 57M parameters"")."
+context_lengths_evaluated,,,not_reported,Input context length (T) is not explicitly stated.
+horizon_lengths_evaluated,96,time steps,stated,Prediction length is 96 for ETT datasets and 24 for ILI (Table 5).
+inference_batch_size,,,not_reported,No inference batch size is mentioned.
diff --git a/result/per_paper/2402.16412/accuracy_efficiency.csv b/result/per_paper/2402.16412/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..9a8d65ac4c71183cb50ca655969f257169659fc9
--- /dev/null
+++ b/result/per_paper/2402.16412/accuracy_efficiency.csv
@@ -0,0 +1,101 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,W,TOTEM,MSE,0.217,
+accuracy,E,TOTEM,MSE,0.193,
+accuracy,T,TOTEM,MSE,0.587,
+accuracy,m1,TOTEM,MSE,0.379,
+accuracy,m2,TOTEM,MSE,0.203,
+accuracy,h1,TOTEM,MSE,0.376,
+accuracy,h2,TOTEM,MSE,0.358,
+accuracy,W,TOTEM,MAE,96,
+accuracy,W,TOTEM,MAE,0.216,
+accuracy,E,TOTEM,MAE,96,
+accuracy,E,TOTEM,MAE,0.264,
+accuracy,T,TOTEM,MAE,96,
+accuracy,T,TOTEM,MAE,0.284,
+accuracy,m1,TOTEM,MAE,96,
+accuracy,m1,TOTEM,MAE,0.384,
+accuracy,m2,TOTEM,MAE,96,
+accuracy,m2,TOTEM,MAE,0.275,
+accuracy,h1,TOTEM,MAE,96,
+accuracy,h1,TOTEM,MAE,0.404,
+accuracy,h2,TOTEM,MAE,96,
+accuracy,h2,TOTEM,MAE,0.345,
+accuracy,N2,TOTEM,MAE,96,
+accuracy,N2,TOTEM,MAE,0.777,
+accuracy,N5,TOTEM,MAE,96,
+accuracy,N5,TOTEM,MAE,0.484,
+accuracy,R,TOTEM,MAE,96,
+accuracy,R,TOTEM,MAE,0.582,
+accuracy,B,TOTEM,MAE,96,
+accuracy,B,TOTEM,MAE,0.739,
+accuracy,S,TOTEM,MAE,96,
+accuracy,S,TOTEM,MAE,0.482,
+accuracy,W,TOTEM,MSE,0.029,0.0012
+accuracy,W,TOTEM,MSE,12.5,
+accuracy,E,TOTEM,MSE,0.065,0.0020
+accuracy,E,TOTEM,MSE,12.5,
+accuracy,m1,TOTEM,MSE,0.041,0.0006
+accuracy,m1,TOTEM,MSE,12.5,
+accuracy,m2,TOTEM,MSE,0.040,0.0020
+accuracy,m2,TOTEM,MSE,12.5,
+accuracy,h1,TOTEM,MSE,0.100,0.0049
+accuracy,h1,TOTEM,MSE,12.5,
+accuracy,h2,TOTEM,MSE,0.075,0.0012
+accuracy,h2,TOTEM,MSE,12.5,
+accuracy,N2,TOTEM,MSE,0.029,0.0015
+accuracy,N2,TOTEM,MSE,12.5,
+accuracy,N5,TOTEM,MSE,0.017,0.0010
+accuracy,N5,TOTEM,MSE,12.5,
+accuracy,R,TOTEM,MSE,0.071,0.0070
+accuracy,R,TOTEM,MSE,12.5,
+accuracy,B,TOTEM,MSE,0.632,0.0087
+accuracy,B,TOTEM,MSE,12.5,
+accuracy,S,TOTEM,MSE,0.057,0.0012
+accuracy,S,TOTEM,MSE,12.5,
+accuracy,F1,TOTEM,F1,79.62,
+accuracy,MSL,TOTEM,F1,82.45,
+accuracy,SMAP,TOTEM,F1,72.88,
+accuracy,SWAT,TOTEM,F1,94.23,
+accuracy,PSM,TOTEM,F1,97.13,
+accuracy,R,TOTEM,F1,76.06,
+accuracy,MSL,TOTEM,F1,82.91,
+accuracy,SMAP,TOTEM,F1,60.95,
+accuracy,SWAT,TOTEM,F1,96.34,
+accuracy,PSM,TOTEM,F1,95.68,
+accuracy,P,TOTEM,F1,83.54,
+accuracy,MSL,TOTEM,F1,82.00,
+accuracy,SMAP,TOTEM,F1,90.60,
+accuracy,SWAT,TOTEM,F1,92.20,
+accuracy,PSM,TOTEM,F1,98.62,
+accuracy,AvgWins,TOTEM,F1,13.3,
+accuracy,W,TOTEM,MSE,0.184,0.0013
+accuracy,E,TOTEM,MSE,0.186,0.0004
+accuracy,T,TOTEM,MSE,0.471,0.0016
+accuracy,m1,TOTEM,MSE,0.328,0.0022
+accuracy,m2,TOTEM,MSE,0.178,0.0000
+accuracy,h1,TOTEM,MSE,0.379,0.0032
+accuracy,h2,TOTEM,MSE,0.295,0.0000
+accuracy,W,TOTEM,"MSE, MAE",0.147,
+accuracy,W,TOTEM,"MSE, MAE",0.195,
+accuracy,W,TOTEM,"MSE, MAE",0.248,
+accuracy,W,TOTEM,"MSE, MAE",0.314,
+accuracy,E,TOTEM,"MSE, MAE",0.135,
+accuracy,E,TOTEM,"MSE, MAE",0.151,
+accuracy,E,TOTEM,"MSE, MAE",0.168,
+accuracy,E,TOTEM,"MSE, MAE",0.200,
+accuracy,T,TOTEM,"MSE, MAE",0.369,
+accuracy,T,TOTEM,"MSE, MAE",0.383,
+accuracy,T,TOTEM,"MSE, MAE",0.397,
+accuracy,T,TOTEM,"MSE, MAE",0.446,
+accuracy,W,TOTEM,"MSE, MAE",0.165,
+accuracy,W,TOTEM,"MSE, MAE",0.207,
+accuracy,W,TOTEM,"MSE, MAE",0.257,
+accuracy,W,TOTEM,"MSE, MAE",0.326,
+accuracy,E,TOTEM,"MSE, MAE",0.178,
+accuracy,E,TOTEM,"MSE, MAE",0.187,
+accuracy,E,TOTEM,"MSE, MAE",0.199,
+accuracy,E,TOTEM,"MSE, MAE",0.236,
+accuracy,T,TOTEM,"MSE, MAE",0.523,
+accuracy,T,TOTEM,"MSE, MAE",0.530,
+accuracy,T,TOTEM,"MSE, MAE",0.549,
+accuracy,T,TOTEM,"MSE, MAE",0.598,
diff --git a/result/per_paper/2402.16412/accuracy_efficiency_traced.csv b/result/per_paper/2402.16412/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ad11c432a0d95a8de5f50619687704258416e827
--- /dev/null
+++ b/result/per_paper/2402.16412/accuracy_efficiency_traced.csv
@@ -0,0 +1,101 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,W,TOTEM,MSE,0.217,,0,2,14
+accuracy,E,TOTEM,MSE,0.193,,0,6,14
+accuracy,T,TOTEM,MSE,0.587,,0,10,14
+accuracy,m1,TOTEM,MSE,0.379,,0,14,14
+accuracy,m2,TOTEM,MSE,0.203,,0,18,14
+accuracy,h1,TOTEM,MSE,0.376,,0,22,14
+accuracy,h2,TOTEM,MSE,0.358,,0,26,14
+accuracy,W,TOTEM,MAE,96,,2,2,1
+accuracy,W,TOTEM,MAE,0.216,,2,2,2
+accuracy,E,TOTEM,MAE,96,,2,3,1
+accuracy,E,TOTEM,MAE,0.264,,2,3,2
+accuracy,T,TOTEM,MAE,96,,2,4,1
+accuracy,T,TOTEM,MAE,0.284,,2,4,2
+accuracy,m1,TOTEM,MAE,96,,2,5,1
+accuracy,m1,TOTEM,MAE,0.384,,2,5,2
+accuracy,m2,TOTEM,MAE,96,,2,6,1
+accuracy,m2,TOTEM,MAE,0.275,,2,6,2
+accuracy,h1,TOTEM,MAE,96,,2,7,1
+accuracy,h1,TOTEM,MAE,0.404,,2,7,2
+accuracy,h2,TOTEM,MAE,96,,2,8,1
+accuracy,h2,TOTEM,MAE,0.345,,2,8,2
+accuracy,N2,TOTEM,MAE,96,,2,10,1
+accuracy,N2,TOTEM,MAE,0.777,,2,10,2
+accuracy,N5,TOTEM,MAE,96,,2,11,1
+accuracy,N5,TOTEM,MAE,0.484,,2,11,2
+accuracy,R,TOTEM,MAE,96,,2,12,1
+accuracy,R,TOTEM,MAE,0.582,,2,12,2
+accuracy,B,TOTEM,MAE,96,,2,13,1
+accuracy,B,TOTEM,MAE,0.739,,2,13,2
+accuracy,S,TOTEM,MAE,96,,2,14,1
+accuracy,S,TOTEM,MAE,0.482,,2,14,2
+accuracy,W,TOTEM,MSE,0.029,0.0012,3,2,2
+accuracy,W,TOTEM,MSE,12.5,,3,2,5
+accuracy,E,TOTEM,MSE,0.065,0.0020,3,6,2
+accuracy,E,TOTEM,MSE,12.5,,3,6,5
+accuracy,m1,TOTEM,MSE,0.041,0.0006,3,10,2
+accuracy,m1,TOTEM,MSE,12.5,,3,10,5
+accuracy,m2,TOTEM,MSE,0.040,0.0020,3,14,2
+accuracy,m2,TOTEM,MSE,12.5,,3,14,5
+accuracy,h1,TOTEM,MSE,0.100,0.0049,3,18,2
+accuracy,h1,TOTEM,MSE,12.5,,3,18,5
+accuracy,h2,TOTEM,MSE,0.075,0.0012,3,22,2
+accuracy,h2,TOTEM,MSE,12.5,,3,22,5
+accuracy,N2,TOTEM,MSE,0.029,0.0015,3,27,2
+accuracy,N2,TOTEM,MSE,12.5,,3,27,5
+accuracy,N5,TOTEM,MSE,0.017,0.0010,3,31,2
+accuracy,N5,TOTEM,MSE,12.5,,3,31,5
+accuracy,R,TOTEM,MSE,0.071,0.0070,3,35,2
+accuracy,R,TOTEM,MSE,12.5,,3,35,5
+accuracy,B,TOTEM,MSE,0.632,0.0087,3,39,2
+accuracy,B,TOTEM,MSE,12.5,,3,39,5
+accuracy,S,TOTEM,MSE,0.057,0.0012,3,43,2
+accuracy,S,TOTEM,MSE,12.5,,3,43,5
+accuracy,F1,TOTEM,F1,79.62,,4,1,2
+accuracy,MSL,TOTEM,F1,82.45,,4,2,2
+accuracy,SMAP,TOTEM,F1,72.88,,4,3,2
+accuracy,SWAT,TOTEM,F1,94.23,,4,4,2
+accuracy,PSM,TOTEM,F1,97.13,,4,5,2
+accuracy,R,TOTEM,F1,76.06,,4,6,2
+accuracy,MSL,TOTEM,F1,82.91,,4,7,2
+accuracy,SMAP,TOTEM,F1,60.95,,4,8,2
+accuracy,SWAT,TOTEM,F1,96.34,,4,9,2
+accuracy,PSM,TOTEM,F1,95.68,,4,10,2
+accuracy,P,TOTEM,F1,83.54,,4,11,2
+accuracy,MSL,TOTEM,F1,82.00,,4,12,2
+accuracy,SMAP,TOTEM,F1,90.60,,4,13,2
+accuracy,SWAT,TOTEM,F1,92.20,,4,14,2
+accuracy,PSM,TOTEM,F1,98.62,,4,15,2
+accuracy,AvgWins,TOTEM,F1,13.3,,4,16,2
+accuracy,W,TOTEM,MSE,0.184,0.0013,5,3,6
+accuracy,E,TOTEM,MSE,0.186,0.0004,5,7,6
+accuracy,T,TOTEM,MSE,0.471,0.0016,5,11,6
+accuracy,m1,TOTEM,MSE,0.328,0.0022,5,15,6
+accuracy,m2,TOTEM,MSE,0.178,0.0000,5,19,6
+accuracy,h1,TOTEM,MSE,0.379,0.0032,5,23,6
+accuracy,h2,TOTEM,MSE,0.295,0.0000,5,27,6
+accuracy,W,TOTEM,"MSE, MAE",0.147,,7,3,2
+accuracy,W,TOTEM,"MSE, MAE",0.195,,7,4,2
+accuracy,W,TOTEM,"MSE, MAE",0.248,,7,5,2
+accuracy,W,TOTEM,"MSE, MAE",0.314,,7,6,2
+accuracy,E,TOTEM,"MSE, MAE",0.135,,7,7,2
+accuracy,E,TOTEM,"MSE, MAE",0.151,,7,8,2
+accuracy,E,TOTEM,"MSE, MAE",0.168,,7,9,2
+accuracy,E,TOTEM,"MSE, MAE",0.200,,7,10,2
+accuracy,T,TOTEM,"MSE, MAE",0.369,,7,11,2
+accuracy,T,TOTEM,"MSE, MAE",0.383,,7,12,2
+accuracy,T,TOTEM,"MSE, MAE",0.397,,7,13,2
+accuracy,T,TOTEM,"MSE, MAE",0.446,,7,14,2
+accuracy,W,TOTEM,"MSE, MAE",0.165,,7,17,2
+accuracy,W,TOTEM,"MSE, MAE",0.207,,7,18,2
+accuracy,W,TOTEM,"MSE, MAE",0.257,,7,19,2
+accuracy,W,TOTEM,"MSE, MAE",0.326,,7,20,2
+accuracy,E,TOTEM,"MSE, MAE",0.178,,7,21,2
+accuracy,E,TOTEM,"MSE, MAE",0.187,,7,22,2
+accuracy,E,TOTEM,"MSE, MAE",0.199,,7,23,2
+accuracy,E,TOTEM,"MSE, MAE",0.236,,7,24,2
+accuracy,T,TOTEM,"MSE, MAE",0.523,,7,25,2
+accuracy,T,TOTEM,"MSE, MAE",0.530,,7,26,2
+accuracy,T,TOTEM,"MSE, MAE",0.549,,7,27,2
+accuracy,T,TOTEM,"MSE, MAE",0.598,,7,28,2
diff --git a/result/per_paper/2402.16412/components_architecture.csv b/result/per_paper/2402.16412/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..611ff888c64dff15054b11572bdfb30195e5e441
--- /dev/null
+++ b/result/per_paper/2402.16412/components_architecture.csv
@@ -0,0 +1,8 @@
+component,what_it_is,provenance,citation,evidence
+VQVAE Architecture,"A vector-quantized variational autoencoder (VQVAE) with a 1D strided CNN encoder, quantizer, latent codebook, and 1D strided transpose CNN decoder.",reused_cited,"Van Den Oord et al., 2017; Esser et al., 2021","The TOTEM VQVAE architecture consists of a 1D strided CNN encoder E, quantizer, latent codebook, and 1D strided transpose CNN decoder D."
+Self-Supervised Tokenization,"A method to discretely tokenize time series data via self-supervised pre-training over a multi-domain corpus, producing a fixed number of discrete tokens encoding univariate waveform shapes.",proposed_here,,TOTEM employs a self-supervised pre-training stage to learn a fixed number of discrete tokens over a multi-domain corpus.
+Discrete Codebook,"A latent codebook storing quantized representations of time series data, frozen after pre-training for use in downstream tasks.",reused_cited,"Van Den Oord et al., 2017","TOTEM's discrete, self-supervised codebook is frozen then leveraged for both in-domain and zero-shot testing across many tasks."
+1D Strided CNN Encoder,A convolutional neural network with strided operations to extract hierarchical features from time series data.,reused_cited,"Zhou et al., 2023; Wu et al., 2022",The TOTEM VQVAE architecture consists of a 1D strided CNN encoder E.
+1D Strided Transpose CNN Decoder,A transposed convolutional neural network with strided operations to reconstruct time series data from latent representations.,reused_cited,"Zhou et al., 2023; Wu et al., 2022",The TOTEM VQVAE architecture consists of a 1D strided transpose CNN decoder D.
+Quantizer,"A module that maps encoder outputs to the nearest latent code in the codebook, enabling discrete representation learning.",reused_cited,"Van Den Oord et al., 2017",The TOTEM VQVAE architecture includes a quantizer.
+Univariate Token Encoding,"A method to encode univariate waveform shapes into discrete tokens, enabling generic multivariate tokenization via stacking.",proposed_here,,TOTEM generically tokenizes multivariate data by stacking collections of univariate tokens.
diff --git a/result/per_paper/2402.16412/computational.csv b/result/per_paper/2402.16412/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..79ecf4fe0b2386cb7a8b9361e9610b59d3040719
--- /dev/null
+++ b/result/per_paper/2402.16412/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware specifications (e.g., GPU/TPU types) are mentioned in the text."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs/TPUs) used for training or inference is not specified."
+training_cost,,,not_reported,"No monetary or computational cost (e.g., cloud credits, energy) is reported for training."
+training_batch_size,,,not_reported,The training batch size is not explicitly stated in the text.
+training_steps_or_epochs,,,not_reported,The number of training steps or epochs is not mentioned.
+precision,,,not_reported,"The numerical precision (e.g., 32-bit, 16-bit) used during training or inference is not specified."
+inference_latency,,,not_reported,"Latency metrics for inference (e.g., time per sample) are not reported."
+inference_throughput,,,not_reported,"Throughput metrics (e.g., samples per second) for inference are not mentioned."
+peak_memory,,,not_reported,Peak memory usage during training or inference is not provided.
+flops_or_macs,,,not_reported,FLOPs or MACs (computational operations) are not discussed.
+num_inference_samples,,,not_reported,The number of samples used during inference experiments is not explicitly stated.
+params,,,not_reported,The total number of parameters in the TOTEM model is not reported.
+context_lengths_evaluated,,,not_reported,"Specific context lengths (e.g., sequence lengths) evaluated during training or testing are not explicitly mentioned."
+horizon_lengths_evaluated,,,not_reported,"Horizon lengths (e.g., forecast/prediction lengths) evaluated are not specified."
+inference_batch_size,,,not_reported,The batch size used during inference is not mentioned.
diff --git a/result/per_paper/2403.00131/accuracy_efficiency.csv b/result/per_paper/2403.00131/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..c581bca7e682a231a1ba8198e2cae95dd88e39ce
--- /dev/null
+++ b/result/per_paper/2403.00131/accuracy_efficiency.csv
@@ -0,0 +1,79 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTml,UniTS-ST,MSE MAE,0.015,
+accuracy,ETThl,UniTS-ST,MSE MAE,0.032,
+accuracy,Electricity,UniTS-ST,MSE MAE,0.031,
+accuracy,Weather,UniTS-ST,MSE MAE,0.025,
+accuracy,Best Count,UniTS-ST,MSE MAE,16,
+accuracy,ETTm1,UniTS-ST,MAE,0.400,
+accuracy,ETTm2,UniTS-ST,MAE,0.291,
+accuracy,ETTh1,UniTS-ST,MAE,0.458,
+accuracy,ETTh2,UniTS-ST,MAE,0.414,
+accuracy,ECL,UniTS-ST,MAE,0.192,
+accuracy,Exchange,UniTS-ST,MAE,0.416,
+accuracy,Traffic,UniTS-ST,MAE,0.620,
+accuracy,Weather,UniTS-ST,MAE,0.259,
+accuracy,Solar-Energy,UniTS-ST,MAE,0.301,
+accuracy,Best Count,UniTS-ST,MAE,0,
+accuracy,Avg.,UniTS-ST,MAE,66.0,
+accuracy,SMD,UniTS-ST,MAE,81.65,
+accuracy,MSL,UniTS-ST,MAE,84.06,
+accuracy,SMAP,UniTS-ST,MAE,69.92,
+accuracy,SWaT,UniTS-ST,MAE,81.43,
+accuracy,PSM,UniTS-ST,MAE,77.10,
+accuracy,Avg.,UniTS-ST,MAE,78.83,
+accuracy,ETTm1,UniTS-ST,MAE,0.036,
+accuracy,ETTh1,UniTS-ST,MAE,0.094,
+accuracy,ECL,UniTS-ST,MAE,0.100,
+accuracy,Weather,UniTS-ST,MAE,0.032,
+accuracy,Best Count,UniTS-ST,MAE,0,
+accuracy,EthanolConcentration,UniTS-ST,accuracy,29.7,
+accuracy,FaceDetection,UniTS-ST,accuracy,67.5,
+accuracy,Handwriting,UniTS-ST,accuracy,26.1,
+accuracy,Heartbeat,UniTS-ST,accuracy,75.1,
+accuracy,JapaneseVowels,UniTS-ST,accuracy,96.2,
+accuracy,PEMS-SF,UniTS-ST,accuracy,88.4,
+accuracy,SelfRegulationSCP1,UniTS-ST,accuracy,89.8,
+accuracy,SelfRegulationSCP2,UniTS-ST,accuracy,51.1,
+accuracy,SpokenArabicDigits,UniTS-ST,accuracy,100.0,
+accuracy,UWaveGestureLibrary,UniTS-ST,accuracy,80.3,
+accuracy,Average Accuracy,UniTS-ST,accuracy,70.4,
+accuracy,HEARTBEAT,UniTS-SUP,accuracy,0.639,
+accuracy,HEARTBEAT,UniTS-PMT,accuracy,0.654,
+accuracy,JAPANESEVOWELS,UniTS-SUP,accuracy,0.922,
+accuracy,JAPANESEVOWELS,UniTS-PMT,accuracy,0.903,
+accuracy,PEMS-SF,UniTS-SUP,accuracy,0.832,
+accuracy,PEMS-SF,UniTS-PMT,accuracy,0.827,
+accuracy,SELFREGULATIONSCP2,UniTS-SUP,accuracy,0.489,
+accuracy,SELFREGULATIONSCP2,UniTS-PMT,accuracy,0.572,
+accuracy,SPOKENARABICDIGITS,UniTS-SUP,accuracy,0.968,
+accuracy,SPOKENARABICDIGITS,UniTS-PMT,accuracy,0.955,
+accuracy,UWAVEGESTURELIBRARY,UniTS-SUP,accuracy,0.822,
+accuracy,UWAVEGESTURELIBRARY,UniTS-PMT,accuracy,0.853,
+accuracy,ECG5000,UniTS-SUP,accuracy,0.928,
+accuracy,ECG5000,UniTS-PMT,accuracy,0.924,
+accuracy,NONINVASIVEFETALECGTHORAX1,UniTS-SUP,accuracy,0.896,
+accuracy,NONINVASIVEFETALECGTHORAX1,UniTS-PMT,accuracy,0.808,
+accuracy,BLINK,UniTS-SUP,accuracy,0.976,
+accuracy,BLINK,UniTS-PMT,accuracy,0.916,
+accuracy,FACEDETECTION,UniTS-SUP,accuracy,0.654,
+accuracy,FACEDETECTION,UniTS-PMT,accuracy,0.58,
+accuracy,ELECTRICDEVICES,UniTS-SUP,accuracy,0.622,
+accuracy,ELECTRICDEVICES,UniTS-PMT,accuracy,0.624,
+accuracy,TRACE,UniTS-SUP,accuracy,0.96,
+accuracy,TRACE,UniTS-PMT,accuracy,0.99,
+accuracy,FORDB,UniTS-SUP,accuracy,0.759,
+accuracy,FORDB,UniTS-PMT,accuracy,0.78,
+accuracy,MOTIONSENSEHAR,UniTS-SUP,accuracy,0.951,
+accuracy,MOTIONSENSEHAR,UniTS-PMT,accuracy,0.958,
+accuracy,EMOPAIN,UniTS-SUP,accuracy,0.797,
+accuracy,EMOPAIN,UniTS-PMT,accuracy,0.814,
+accuracy,CHINATOWN,UniTS-SUP,accuracy,0.98,
+accuracy,CHINATOWN,UniTS-PMT,accuracy,0.98,
+accuracy,MELBOURNEPEDESTRIAN,UniTS-SUP,accuracy,0.876,
+accuracy,MELBOURNEPEDESTRIAN,UniTS-PMT,accuracy,0.839,
+accuracy,SHAREPRICEINCREASE,UniTS-SUP,accuracy,0.618,
+accuracy,SHAREPRICEINCREASE,UniTS-PMT,accuracy,0.638,
+accuracy,BEST COUNT,UniTS-SUP,accuracy,3,
+accuracy,BEST COUNT,UniTS-PMT,accuracy,7,
+accuracy,AVERAGE SCORE,UniTS-SUP,accuracy,0.816,
+accuracy,AVERAGE SCORE,UniTS-PMT,accuracy,0.812,
diff --git a/result/per_paper/2403.00131/accuracy_efficiency_traced.csv b/result/per_paper/2403.00131/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a5f4722b3ddacdfaec5fb21d90d4640877835d38
--- /dev/null
+++ b/result/per_paper/2403.00131/accuracy_efficiency_traced.csv
@@ -0,0 +1,79 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTml,UniTS-ST,MSE MAE,0.015,,1,2,2
+accuracy,ETThl,UniTS-ST,MSE MAE,0.032,,1,7,2
+accuracy,Electricity,UniTS-ST,MSE MAE,0.031,,1,12,2
+accuracy,Weather,UniTS-ST,MSE MAE,0.025,,1,17,2
+accuracy,Best Count,UniTS-ST,MSE MAE,16,,1,22,2
+accuracy,ETTm1,UniTS-ST,MAE,0.400,,3,2,13
+accuracy,ETTm2,UniTS-ST,MAE,0.291,,3,3,13
+accuracy,ETTh1,UniTS-ST,MAE,0.458,,3,4,13
+accuracy,ETTh2,UniTS-ST,MAE,0.414,,3,5,13
+accuracy,ECL,UniTS-ST,MAE,0.192,,3,6,13
+accuracy,Exchange,UniTS-ST,MAE,0.416,,3,7,13
+accuracy,Traffic,UniTS-ST,MAE,0.620,,3,8,13
+accuracy,Weather,UniTS-ST,MAE,0.259,,3,9,13
+accuracy,Solar-Energy,UniTS-ST,MAE,0.301,,3,10,13
+accuracy,Best Count,UniTS-ST,MAE,0,,3,11,13
+accuracy,Avg.,UniTS-ST,MAE,66.0,,3,14,13
+accuracy,SMD,UniTS-ST,MAE,81.65,,3,16,13
+accuracy,MSL,UniTS-ST,MAE,84.06,,3,17,13
+accuracy,SMAP,UniTS-ST,MAE,69.92,,3,18,13
+accuracy,SWaT,UniTS-ST,MAE,81.43,,3,19,13
+accuracy,PSM,UniTS-ST,MAE,77.10,,3,20,13
+accuracy,Avg.,UniTS-ST,MAE,78.83,,3,21,13
+accuracy,ETTm1,UniTS-ST,MAE,0.036,,3,24,13
+accuracy,ETTh1,UniTS-ST,MAE,0.094,,3,25,13
+accuracy,ECL,UniTS-ST,MAE,0.100,,3,26,13
+accuracy,Weather,UniTS-ST,MAE,0.032,,3,27,13
+accuracy,Best Count,UniTS-ST,MAE,0,,3,28,13
+accuracy,EthanolConcentration,UniTS-ST,accuracy,29.7,,5,3,18
+accuracy,FaceDetection,UniTS-ST,accuracy,67.5,,5,4,18
+accuracy,Handwriting,UniTS-ST,accuracy,26.1,,5,5,18
+accuracy,Heartbeat,UniTS-ST,accuracy,75.1,,5,6,18
+accuracy,JapaneseVowels,UniTS-ST,accuracy,96.2,,5,7,18
+accuracy,PEMS-SF,UniTS-ST,accuracy,88.4,,5,8,18
+accuracy,SelfRegulationSCP1,UniTS-ST,accuracy,89.8,,5,9,18
+accuracy,SelfRegulationSCP2,UniTS-ST,accuracy,51.1,,5,10,18
+accuracy,SpokenArabicDigits,UniTS-ST,accuracy,100.0,,5,11,18
+accuracy,UWaveGestureLibrary,UniTS-ST,accuracy,80.3,,5,12,18
+accuracy,Average Accuracy,UniTS-ST,accuracy,70.4,,5,13,18
+accuracy,HEARTBEAT,UniTS-SUP,accuracy,0.639,,7,1,1
+accuracy,HEARTBEAT,UniTS-PMT,accuracy,0.654,,7,1,2
+accuracy,JAPANESEVOWELS,UniTS-SUP,accuracy,0.922,,7,2,1
+accuracy,JAPANESEVOWELS,UniTS-PMT,accuracy,0.903,,7,2,2
+accuracy,PEMS-SF,UniTS-SUP,accuracy,0.832,,7,3,1
+accuracy,PEMS-SF,UniTS-PMT,accuracy,0.827,,7,3,2
+accuracy,SELFREGULATIONSCP2,UniTS-SUP,accuracy,0.489,,7,4,1
+accuracy,SELFREGULATIONSCP2,UniTS-PMT,accuracy,0.572,,7,4,2
+accuracy,SPOKENARABICDIGITS,UniTS-SUP,accuracy,0.968,,7,5,1
+accuracy,SPOKENARABICDIGITS,UniTS-PMT,accuracy,0.955,,7,5,2
+accuracy,UWAVEGESTURELIBRARY,UniTS-SUP,accuracy,0.822,,7,6,1
+accuracy,UWAVEGESTURELIBRARY,UniTS-PMT,accuracy,0.853,,7,6,2
+accuracy,ECG5000,UniTS-SUP,accuracy,0.928,,7,7,1
+accuracy,ECG5000,UniTS-PMT,accuracy,0.924,,7,7,2
+accuracy,NONINVASIVEFETALECGTHORAX1,UniTS-SUP,accuracy,0.896,,7,8,1
+accuracy,NONINVASIVEFETALECGTHORAX1,UniTS-PMT,accuracy,0.808,,7,8,2
+accuracy,BLINK,UniTS-SUP,accuracy,0.976,,7,9,1
+accuracy,BLINK,UniTS-PMT,accuracy,0.916,,7,9,2
+accuracy,FACEDETECTION,UniTS-SUP,accuracy,0.654,,7,10,1
+accuracy,FACEDETECTION,UniTS-PMT,accuracy,0.58,,7,10,2
+accuracy,ELECTRICDEVICES,UniTS-SUP,accuracy,0.622,,7,11,1
+accuracy,ELECTRICDEVICES,UniTS-PMT,accuracy,0.624,,7,11,2
+accuracy,TRACE,UniTS-SUP,accuracy,0.96,,7,12,1
+accuracy,TRACE,UniTS-PMT,accuracy,0.99,,7,12,2
+accuracy,FORDB,UniTS-SUP,accuracy,0.759,,7,13,1
+accuracy,FORDB,UniTS-PMT,accuracy,0.78,,7,13,2
+accuracy,MOTIONSENSEHAR,UniTS-SUP,accuracy,0.951,,7,14,1
+accuracy,MOTIONSENSEHAR,UniTS-PMT,accuracy,0.958,,7,14,2
+accuracy,EMOPAIN,UniTS-SUP,accuracy,0.797,,7,15,1
+accuracy,EMOPAIN,UniTS-PMT,accuracy,0.814,,7,15,2
+accuracy,CHINATOWN,UniTS-SUP,accuracy,0.98,,7,16,1
+accuracy,CHINATOWN,UniTS-PMT,accuracy,0.98,,7,16,2
+accuracy,MELBOURNEPEDESTRIAN,UniTS-SUP,accuracy,0.876,,7,17,1
+accuracy,MELBOURNEPEDESTRIAN,UniTS-PMT,accuracy,0.839,,7,17,2
+accuracy,SHAREPRICEINCREASE,UniTS-SUP,accuracy,0.618,,7,18,1
+accuracy,SHAREPRICEINCREASE,UniTS-PMT,accuracy,0.638,,7,18,2
+accuracy,BEST COUNT,UniTS-SUP,accuracy,3,,7,19,1
+accuracy,BEST COUNT,UniTS-PMT,accuracy,7,,7,19,2
+accuracy,AVERAGE SCORE,UniTS-SUP,accuracy,0.816,,7,20,1
+accuracy,AVERAGE SCORE,UniTS-PMT,accuracy,0.812,,7,20,2
diff --git a/result/per_paper/2403.00131/components_architecture.csv b/result/per_paper/2403.00131/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2e4c5a8fcfc938ab3f2d20ed6be915fc2d65dd66
--- /dev/null
+++ b/result/per_paper/2403.00131/components_architecture.csv
@@ -0,0 +1,8 @@
+component,what_it_is,provenance,citation,evidence
+Task Tokenization,"Encodes task specifications into a unified token representation, enabling universal task specification without post-hoc architectural modifications.",proposed_here,,"UNITS encodes task specifications into a unified token representation, enabling universal task specification without post-hoc architectural modifications."
+Modified Transformer Block,"A transformer block adapted to capture universal time series representations, enabling transferability across diverse domains and temporal scales.",proposed_here,,"UNITS employs a modified transformer block to capture universal time series representations, enabling transferability from a heterogeneous, multi-domain pre-training dataset."
+Dynamic Linear Operator,"A module to model complex relationships between data points along the time dimension, enhancing adaptability to temporal dynamics.",proposed_here,,We introduce a dynamic linear operator to model complex relationships between data points along the time dimension.
+Feature Space Interference Reduction Module,"A module to reduce interference in the feature space of heterogeneous data, improving performance on multi-domain tasks.",proposed_here,,We introduce a module to reduce interference in the feature space of heterogeneous data.
+Unified Time Series Architecture,A framework that processes heterogeneous time series data with varying numbers of variables and sequence lengths without altering its network structure.,proposed_here,,UNITS processes heterogeneous time series data with varying numbers of variables and sequence lengths without altering its network structure.
+Self-Attention Mechanism,A mechanism that applies self-attention across time and variable dimensions to adapt to diverse temporal dynamics.,proposed_here,,UNITS employs self-attention across time and variable dimensions to adapt to diverse temporal dynamics.
+Masked Reconstruction Pre-training,A pre-training approach that enables joint optimization for generative and predictive tasks by reconstructing masked time series data.,proposed_here,,"We use a masked reconstruction pre-training approach, enabling UNITS to be jointly optimized for generative and predictive tasks."
diff --git a/result/per_paper/2403.00131/computational.csv b/result/per_paper/2403.00131/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2403.00131/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2403.07815/accuracy_efficiency.csv b/result/per_paper/2403.07815/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5545dfb20716648fb28448f2911177eac9524d89
--- /dev/null
+++ b/result/per_paper/2403.07815/accuracy_efficiency.csv
@@ -0,0 +1,415 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Australian Electricity,Chronos-T3 (Large),WQL,1.333,
+accuracy,Australian Electricity,Chronos-T3 (Base),WQL,1.319,
+accuracy,Australian Electricity,Chronos-T3 (Base),WQL,1.399,
+accuracy,Australian Electricity,Chronos-T3 (MHI),WQL,1.114,
+accuracy,Australian Electricity,Chronos-GPT2,WQL,1.310,
+accuracy,Car,Chronos-T3 (Large),WQL,0.906,
+accuracy,Car,Chronos-T3 (Base),WQL,0.899,
+accuracy,Car,Chronos-T3 (Base),WQL,0.887,
+accuracy,Car,Chronos-T3 (MHI),WQL,0.893,
+accuracy,Car,Chronos-GPT2,WQL,0.881,
+accuracy,CIF 2016,Chronos-T3 (Large),WQL,0.986,
+accuracy,CIF 2016,Chronos-T3 (Base),WQL,0.981,
+accuracy,CIF 2016,Chronos-T3 (Base),WQL,0.989,
+accuracy,CIF 2016,Chronos-T3 (MHI),WQL,1.053,
+accuracy,CIF 2016,Chronos-GPT2,WQL,1.046,
+accuracy,Covid Deaths,Chronos-T3 (Large),WQL,42.550,
+accuracy,Covid Deaths,Chronos-T3 (Base),WQL,42.687,
+accuracy,Covid Deaths,Chronos-T3 (Base),WQL,42.670,
+accuracy,Covid Deaths,Chronos-T3 (MHI),WQL,43.621,
+accuracy,Covid Deaths,Chronos-GPT2,WQL,48.215,
+accuracy,Dominick,Chronos-T3 (Large),WQL,0.818,
+accuracy,Dominick,Chronos-T3 (Base),WQL,0.816,
+accuracy,Dominick,Chronos-T3 (Base),WQL,0.819,
+accuracy,Dominick,Chronos-T3 (MHI),WQL,0.833,
+accuracy,Dominick,Chronos-GPT2,WQL,0.820,
+accuracy,ERCOT Load,Chronos-T3 (Large),WQL,0.617,
+accuracy,ERCOT Load,Chronos-T3 (Base),WQL,0.550,
+accuracy,ERCOT Load,Chronos-T3 (Base),WQL,0.573,
+accuracy,ERCOT Load,Chronos-T3 (MHI),WQL,0.588,
+accuracy,ERCOT Load,Chronos-GPT2,WQL,0.561,
+accuracy,ETT (15 Min.),Chronos-T3 (Large),WQL,0.741,
+accuracy,ETT (15 Min.),Chronos-T3 (Base),WQL,0.739,
+accuracy,ETT (15 Min.),Chronos-T3 (Base),WQL,0.710,
+accuracy,ETT (15 Min.),Chronos-T3 (MHI),WQL,0.790,
+accuracy,ETT (15 Min.),Chronos-GPT2,WQL,0.796,
+accuracy,ETT (Hourly),Chronos-T3 (Large),WQL,0.735,
+accuracy,ETT (Hourly),Chronos-T3 (Base),WQL,0.789,
+accuracy,ETT (Hourly),Chronos-T3 (Base),WQL,0.789,
+accuracy,ETT (Hourly),Chronos-T3 (MHI),WQL,0.797,
+accuracy,ETT (Hourly),Chronos-GPT2,WQL,0.768,
+accuracy,Exchange Rate,Chronos-T3 (Large),WQL,2.975,
+accuracy,Exchange Rate,Chronos-T3 (Base),WQL,2.433,
+accuracy,Exchange Rate,Chronos-T3 (Base),WQL,2.252,
+accuracy,Exchange Rate,Chronos-T3 (MHI),WQL,2.030,
+accuracy,Exchange Rate,Chronos-GPT2,WQL,2.335,
+accuracy,FRED-MD,Chronos-T3 (Large),WQL,0.500,
+accuracy,FRED-MD,Chronos-T3 (Base),WQL,0.486,
+accuracy,FRED-MD,Chronos-T3 (Base),WQL,0.496,
+accuracy,FRED-MD,Chronos-T3 (MHI),WQL,0.483,
+accuracy,FRED-MD,Chronos-GPT2,WQL,0.468,
+accuracy,Hospital,Chronos-T3 (Large),WQL,0.810,
+accuracy,Hospital,Chronos-T3 (Base),WQL,0.810,
+accuracy,Hospital,Chronos-T3 (Base),WQL,0.815,
+accuracy,Hospital,Chronos-T3 (MHI),WQL,0.817,
+accuracy,Hospital,Chronos-GPT2,WQL,0.831,
+accuracy,MI (Monthly),Chronos-T3 (Large),WQL,1.090,
+accuracy,MI (Monthly),Chronos-T3 (Base),WQL,1.117,
+accuracy,MI (Monthly),Chronos-T3 (Base),WQL,1.169,
+accuracy,MI (Monthly),Chronos-T3 (MHI),WQL,1.174,
+accuracy,MI (Monthly),Chronos-GPT2,WQL,1.182,
+accuracy,MI (Quarterly),Chronos-T3 (Large),WQL,1.713,
+accuracy,MI (Quarterly),Chronos-T3 (Base),WQL,1.739,
+accuracy,MI (Quarterly),Chronos-T3 (Base),WQL,1.764,
+accuracy,MI (Quarterly),Chronos-T3 (MHI),WQL,1.785,
+accuracy,MI (Quarterly),Chronos-GPT2,WQL,1.785,
+accuracy,MI (Yearly),Chronos-T3 (Large),WQL,4.301,
+accuracy,MI (Yearly),Chronos-T3 (Base),WQL,4.624,
+accuracy,MI (Yearly),Chronos-T3 (Base),WQL,4.659,
+accuracy,MI (Yearly),Chronos-T3 (MHI),WQL,4.956,
+accuracy,MI (Yearly),Chronos-GPT2,WQL,4.751,
+accuracy,M3 (Monthly),Chronos-T3 (Large),WQL,0.857,
+accuracy,M3 (Monthly),Chronos-T3 (Base),WQL,0.868,
+accuracy,M3 (Monthly),Chronos-T3 (Base),WQL,0.885,
+accuracy,M3 (Monthly),Chronos-T3 (MHI),WQL,0.900,
+accuracy,M3 (Monthly),Chronos-GPT2,WQL,0.930,
+accuracy,M3 (Quarterly),Chronos-T3 (Large),WQL,1.181,
+accuracy,M3 (Quarterly),Chronos-T3 (Base),WQL,1.199,
+accuracy,M3 (Quarterly),Chronos-T3 (Base),WQL,1.256,
+accuracy,M3 (Quarterly),Chronos-T3 (MHI),WQL,1.289,
+accuracy,M3 (Quarterly),Chronos-GPT2,WQL,1.241,
+accuracy,M3 (Yearly),Chronos-T3 (Large),WQL,3.106,
+accuracy,M3 (Yearly),Chronos-T3 (Base),WQL,3.269,
+accuracy,M3 (Yearly),Chronos-T3 (Base),WQL,3.276,
+accuracy,M3 (Yearly),Chronos-T3 (MHI),WQL,3.305,
+accuracy,M3 (Yearly),Chronos-GPT2,WQL,3.158,
+accuracy,M4 (Quarterly),Chronos-T3 (Large),WQL,1.216,
+accuracy,M4 (Quarterly),Chronos-T3 (Base),WQL,1.231,
+accuracy,M4 (Quarterly),Chronos-T3 (Base),WQL,1.246,
+accuracy,M4 (Quarterly),Chronos-T3 (MHI),WQL,1.271,
+accuracy,M4 (Quarterly),Chronos-GPT2,WQL,1.312,
+accuracy,M4 (Yearly),Chronos-T3 (Large),WQL,3.606,
+accuracy,M4 (Yearly),Chronos-T3 (Base),WQL,3.678,
+accuracy,M4 (Yearly),Chronos-T3 (Base),WQL,3.651,
+accuracy,M4 (Yearly),Chronos-T3 (MHI),WQL,3.743,
+accuracy,M4 (Yearly),Chronos-GPT2,WQL,3.933,
+accuracy,M5,Chronos-T3 (Large),WQL,0.944,
+accuracy,M5,Chronos-T3 (Base),WQL,0.939,
+accuracy,M5,Chronos-T3 (Base),WQL,0.940,
+accuracy,M5,Chronos-T3 (MHI),WQL,0.944,
+accuracy,M5,Chronos-GPT2,WQL,0.969,
+accuracy,NNS (Daily),Chronos-T3 (Large),WQL,0.573,
+accuracy,NNS (Daily),Chronos-T3 (Base),WQL,0.585,
+accuracy,NNS (Daily),Chronos-T3 (Base),WQL,0.615,
+accuracy,NNS (Daily),Chronos-T3 (MHI),WQL,0.642,
+accuracy,NNS (Daily),Chronos-GPT2,WQL,0.661,
+accuracy,NNS (Weekly),Chronos-T3 (Large),WQL,0.940,
+accuracy,NNS (Weekly),Chronos-T3 (Base),WQL,0.938,
+accuracy,NNS (Weekly),Chronos-T3 (Base),WQL,0.944,
+accuracy,NNS (Weekly),Chronos-T3 (MHI),WQL,0.947,
+accuracy,NNS (Weekly),Chronos-GPT2,WQL,0.963,
+accuracy,Tourism (Monthly),Chronos-T3 (Large),WQL,1.761,
+accuracy,Tourism (Monthly),Chronos-T3 (Base),WQL,1.828,
+accuracy,Tourism (Monthly),Chronos-T3 (Base),WQL,1.900,
+accuracy,Tourism (Monthly),Chronos-T3 (MHI),WQL,1.950,
+accuracy,Tourism (Monthly),Chronos-GPT2,WQL,1.783,
+accuracy,Tourism (Quarterly),Chronos-T3 (Large),WQL,1.677,
+accuracy,Tourism (Quarterly),Chronos-T3 (Base),WQL,1.717,
+accuracy,Tourism (Quarterly),Chronos-T3 (Base),WQL,1.730,
+accuracy,Tourism (Quarterly),Chronos-T3 (MHI),WQL,1.829,
+accuracy,Tourism (Quarterly),Chronos-GPT2,WQL,1.828,
+accuracy,Tourism (Yearly),Chronos-T3 (Large),WQL,3.755,
+accuracy,Tourism (Yearly),Chronos-T3 (Base),WQL,3.900,
+accuracy,Tourism (Yearly),Chronos-T3 (Base),WQL,3.901,
+accuracy,Tourism (Yearly),Chronos-T3 (MHI),WQL,4.048,
+accuracy,Tourism (Yearly),Chronos-GPT2,WQL,3.882,
+accuracy,Traffic,Chronos-T3 (Large),WQL,0.804,
+accuracy,Traffic,Chronos-T3 (Base),WQL,0.828,
+accuracy,Traffic,Chronos-T3 (Base),WQL,0.837,
+accuracy,Traffic,Chronos-T3 (MHI),WQL,0.850,
+accuracy,Traffic,Chronos-GPT2,WQL,0.818,
+accuracy,Weather,Chronos-T3 (Large),WQL,0.822,
+accuracy,Weather,Chronos-T3 (Base),WQL,0.824,
+accuracy,Weather,Chronos-T3 (Base),WQL,0.836,
+accuracy,Weather,Chronos-T3 (MHI),WQL,0.853,
+accuracy,Weather,Chronos-GPT2,WQL,0.858,
+accuracy,Avg. Relative Score,Chronos-T3 (Large),WQL,0.823,
+accuracy,Avg. Relative Score,Chronos-T3 (Base),WQL,0.832,
+accuracy,Avg. Relative Score,Chronos-T3 (Base),WQL,0.841,
+accuracy,Avg. Relative Score,Chronos-T3 (MHI),WQL,0.850,
+accuracy,Avg. Relative Score,Chronos-GPT2,WQL,0.852,
+accuracy,Avg. Rank,Chronos-T3 (Large),WQL,5.351,
+accuracy,Avg. Rank,Chronos-T3 (Base),WQL,9.296,
+accuracy,Avg. Rank,Chronos-T3 (Base),WQL,10.593,
+accuracy,Avg. Rank,Chronos-T3 (MHI),WQL,12.037,
+accuracy,Avg. Rank,Chronos-GPT2,WQL,11.630,
+accuracy,Australian Electricity,Chronos-T5(Large),WQL,0.067,
+accuracy,Australian Electricity,Chronos-T5(Basic),WQL,0.075,
+accuracy,Australian Electricity,Chronos-T5(Small),WQL,0.074,
+accuracy,Australian Electricity,Chronos-T5(Mini),WQL,0.063,
+accuracy,Car Parts,Chronos-T5(Large),WQL,1.060,
+accuracy,Car Parts,Chronos-T5(Basic),WQL,1.057,
+accuracy,Car Parts,Chronos-T5(Small),WQL,1.029,
+accuracy,Car Parts,Chronos-T5(Mini),WQL,1.024,
+accuracy,CIF 2016,Chronos-T5(Large),WQL,0.014,
+accuracy,CIF 2016,Chronos-T5(Basic),WQL,0.013,
+accuracy,CIF 2016,Chronos-T5(Small),WQL,0.015,
+accuracy,CIF 2016,Chronos-T5(Mini),WQL,0.013,
+accuracy,Covid Deaths,Chronos-T5(Large),WQL,0.045,
+accuracy,Covid Deaths,Chronos-T5(Basic),WQL,0.048,
+accuracy,Covid Deaths,Chronos-T5(Small),WQL,0.059,
+accuracy,Covid Deaths,Chronos-T5(Mini),WQL,0.084,
+accuracy,Donnick,Chronos-T5(Large),WQL,0.332,
+accuracy,Donnick,Chronos-T5(Basic),WQL,0.333,
+accuracy,Donnick,Chronos-T5(Small),WQL,0.338,
+accuracy,Donnick,Chronos-T5(Mini),WQL,0.346,
+accuracy,ERCOT Load,Chronos-T5(Large),WQL,0.019,
+accuracy,ERCOT Load,Chronos-T5(Basic),WQL,0.016,
+accuracy,ERCOT Load,Chronos-T5(Small),WQL,0.018,
+accuracy,ERCOT Load,Chronos-T5(Mini),WQL,0.018,
+accuracy,ETT (15 Min.),Chronos-T5(Large),WQL,0.068,
+accuracy,ETT (15 Min.),Chronos-T5(Basic),WQL,0.069,
+accuracy,ETT (15 Min.),Chronos-T5(Small),WQL,0.064,
+accuracy,ETT (15 Min.),Chronos-T5(Mini),WQL,0.072,
+accuracy,ETT (Hourly),Chronos-T5(Large),WQL,0.073,
+accuracy,ETT (Hourly),Chronos-T5(Basic),WQL,0.081,
+accuracy,ETT (Hourly),Chronos-T5(Small),WQL,0.080,
+accuracy,ETT (Hourly),Chronos-T5(Mini),WQL,0.085,
+accuracy,Exchange Rate,Chronos-T5(Large),WQL,0.013,
+accuracy,Exchange Rate,Chronos-T5(Basic),WQL,0.014,
+accuracy,Exchange Rate,Chronos-T5(Small),WQL,0.013,
+accuracy,Exchange Rate,Chronos-T5(Mini),WQL,0.012,
+accuracy,FRED-MD,Chronos-T5(Large),WQL,0.020,
+accuracy,FRED-MD,Chronos-T5(Basic),WQL,0.022,
+accuracy,FRED-MD,Chronos-T5(Small),WQL,0.017,
+accuracy,FRED-MD,Chronos-T5(Mini),WQL,0.017,
+accuracy,Hospital,Chronos-T5(Large),WQL,0.056,
+accuracy,Hospital,Chronos-T5(Basic),WQL,0.056,
+accuracy,Hospital,Chronos-T5(Small),WQL,0.057,
+accuracy,Hospital,Chronos-T5(Mini),WQL,0.058,
+accuracy,M1 (Monthly),Chronos-T5(Large),WQL,0.130,
+accuracy,M1 (Monthly),Chronos-T5(Basic),WQL,0.128,
+accuracy,M1 (Monthly),Chronos-T5(Small),WQL,0.139,
+accuracy,M1 (Monthly),Chronos-T5(Mini),WQL,0.138,
+accuracy,M1 (Quarterly),Chronos-T5(Large),WQL,0.107,
+accuracy,M1 (Quarterly),Chronos-T5(Basic),WQL,0.105,
+accuracy,M1 (Quarterly),Chronos-T5(Small),WQL,0.103,
+accuracy,M1 (Quarterly),Chronos-T5(Mini),WQL,0.103,
+accuracy,M1 (Yearly),Chronos-T5(Large),WQL,0.183,
+accuracy,M1 (Yearly),Chronos-T5(Basic),WQL,0.181,
+accuracy,M1 (Yearly),Chronos-T5(Small),WQL,0.172,
+accuracy,M1 (Yearly),Chronos-T5(Mini),WQL,0.179,
+accuracy,M3 (Monthly),Chronos-T5(Large),WQL,0.096,
+accuracy,M3 (Monthly),Chronos-T5(Basic),WQL,0.097,
+accuracy,M3 (Monthly),Chronos-T5(Small),WQL,0.100,
+accuracy,M3 (Monthly),Chronos-T5(Mini),WQL,0.099,
+accuracy,M3 (Quarterly),Chronos-T5(Large),WQL,0.074,
+accuracy,M3 (Quarterly),Chronos-T5(Basic),WQL,0.076,
+accuracy,M3 (Quarterly),Chronos-T5(Small),WQL,0.079,
+accuracy,M3 (Quarterly),Chronos-T5(Mini),WQL,0.081,
+accuracy,M3 (Yearly),Chronos-T5(Large),WQL,0.151,
+accuracy,M3 (Yearly),Chronos-T5(Basic),WQL,0.153,
+accuracy,M3 (Yearly),Chronos-T5(Small),WQL,0.155,
+accuracy,M3 (Yearly),Chronos-T5(Mini),WQL,0.159,
+accuracy,M4 (Quarterly),Chronos-T5(Large),WQL,0.082,
+accuracy,M4 (Quarterly),Chronos-T5(Basic),WQL,0.083,
+accuracy,M4 (Quarterly),Chronos-T5(Small),WQL,0.084,
+accuracy,M4 (Quarterly),Chronos-T5(Mini),WQL,0.086,
+accuracy,M4 (Yearly),Chronos-T5(Large),WQL,0.134,
+accuracy,M4 (Yearly),Chronos-T5(Basic),WQL,0.137,
+accuracy,M4 (Yearly),Chronos-T5(Small),WQL,0.136,
+accuracy,M4 (Yearly),Chronos-T5(Mini),WQL,0.140,
+accuracy,M5,Chronos-T5(Large),WQL,0.587,
+accuracy,M5,Chronos-T5(Basic),WQL,0.586,
+accuracy,M5,Chronos-T5(Small),WQL,0.590,
+accuracy,M5,Chronos-T5(Mini),WQL,0.595,
+accuracy,NN5 (Daily),Chronos-T5(Large),WQL,0.156,
+accuracy,NN5 (Daily),Chronos-T5(Basic),WQL,0.161,
+accuracy,NN5 (Daily),Chronos-T5(Small),WQL,0.169,
+accuracy,NN5 (Daily),Chronos-T5(Mini),WQL,0.173,
+accuracy,NN5 (Weekly),Chronos-T5(Large),WQL,0.091,
+accuracy,NN5 (Weekly),Chronos-T5(Basic),WQL,0.091,
+accuracy,NN5 (Weekly),Chronos-T5(Small),WQL,0.090,
+accuracy,NN5 (Weekly),Chronos-T5(Mini),WQL,0.091,
+accuracy,Tourism (Monthly),Chronos-T5(Large),WQL,0.100,
+accuracy,Tourism (Monthly),Chronos-T5(Basic),WQL,0.103,
+accuracy,Tourism (Monthly),Chronos-T5(Small),WQL,0.113,
+accuracy,Tourism (Monthly),Chronos-T5(Mini),WQL,0.109,
+accuracy,Tourism (Quarterly),Chronos-T5(Large),WQL,0.061,
+accuracy,Tourism (Quarterly),Chronos-T5(Basic),WQL,0.069,
+accuracy,Tourism (Quarterly),Chronos-T5(Small),WQL,0.069,
+accuracy,Tourism (Quarterly),Chronos-T5(Mini),WQL,0.074,
+accuracy,Tourism (Yearly),Chronos-T5(Large),WQL,0.183,
+accuracy,Tourism (Yearly),Chronos-T5(Basic),WQL,0.207,
+accuracy,Tourism (Yearly),Chronos-T5(Small),WQL,0.200,
+accuracy,Tourism (Yearly),Chronos-T5(Mini),WQL,0.218,
+accuracy,Traffic,Chronos-T5(Large),WQL,0.256,
+accuracy,Traffic,Chronos-T5(Basic),WQL,0.264,
+accuracy,Traffic,Chronos-T5(Small),WQL,0.263,
+accuracy,Traffic,Chronos-T5(Mini),WQL,0.264,
+accuracy,Weather,Chronos-T5(Large),WQL,0.139,
+accuracy,Weather,Chronos-T5(Basic),WQL,0.140,
+accuracy,Weather,Chronos-T5(Small),WQL,0.143,
+accuracy,Weather,Chronos-T5(Mini),WQL,0.150,
+accuracy,Agg. Relative Score,Chronos-T5(Large),WQL,0.645,
+accuracy,Agg. Relative Score,Chronos-T5(Basic),WQL,0.662,
+accuracy,Agg. Relative Score,Chronos-T5(Small),WQL,0.667,
+accuracy,Agg. Relative Score,Chronos-T5(Mini),WQL,0.678,
+accuracy,Avg. Rank,Chronos-T5(Large),WQL,5.333,
+accuracy,Avg. Rank,Chronos-T5(Basic),WQL,9.407,
+accuracy,Avg. Rank,Chronos-T5(Small),WQL,9.889,
+accuracy,Avg. Rank,Chronos-T5(Mini),WQL,11.296,
+accuracy,Electricity (15 Min.),Chronos-T5 (Large),MASE,0.391,
+accuracy,Electricity (15 Min.),Chronos-T5 (Base),MASE,0.394,
+accuracy,Electricity (15 Min.),Chronos-T5 (Small),MASE,0.418,
+accuracy,Electricity (15 Min.),Chronos-T5 (Mid),MASE,0.445,
+accuracy,Electricity (15 Min.),Chronos-GPT2,MASE,0.388,
+accuracy,Electricity (Hourly),Chronos-T5 (Large),MASE,1.439,
+accuracy,Electricity (Hourly),Chronos-T5 (Base),MASE,1.590,
+accuracy,Electricity (Hourly),Chronos-T5 (Small),MASE,1.477,
+accuracy,Electricity (Hourly),Chronos-T5 (Mid),MASE,1.348,
+accuracy,Electricity (Hourly),Chronos-GPT2,MASE,1.636,
+accuracy,Electricity (Weekly),Chronos-T5 (Large),MASE,1.739,
+accuracy,Electricity (Weekly),Chronos-T5 (Base),MASE,1.801,
+accuracy,Electricity (Weekly),Chronos-T5 (Small),MASE,1.942,
+accuracy,Electricity (Weekly),Chronos-T5 (Mid),MASE,1.954,
+accuracy,Electricity (Weekly),Chronos-GPT2,MASE,1.720,
+accuracy,KDD Cap 2018,Chronos-T5 (Large),MASE,0.683,
+accuracy,KDD Cap 2018,Chronos-T5 (Base),MASE,0.646,
+accuracy,KDD Cap 2018,Chronos-T5 (Small),MASE,0.687,
+accuracy,KDD Cap 2018,Chronos-T5 (Mid),MASE,0.667,
+accuracy,KDD Cap 2018,Chronos-GPT2,MASE,0.881,
+accuracy,London Smart Meters,Chronos-T5 (Large),MASE,0.828,
+accuracy,London Smart Meters,Chronos-T5 (Base),MASE,0.838,
+accuracy,London Smart Meters,Chronos-T5 (Small),MASE,0.846,
+accuracy,London Smart Meters,Chronos-T5 (Mid),MASE,0.857,
+accuracy,London Smart Meters,Chronos-GPT2,MASE,0.842,
+accuracy,M4 (Daily),Chronos-T5 (Large),MASE,3.144,
+accuracy,M4 (Daily),Chronos-T5 (Base),MASE,3.160,
+accuracy,M4 (Daily),Chronos-T5 (Small),MASE,3.148,
+accuracy,M4 (Daily),Chronos-T5 (Mid),MASE,3.154,
+accuracy,M4 (Daily),Chronos-GPT2,MASE,3.079,
+accuracy,M4 (Hourly),Chronos-T5 (Large),MASE,0.682,
+accuracy,M4 (Hourly),Chronos-T5 (Base),MASE,0.694,
+accuracy,M4 (Hourly),Chronos-T5 (Small),MASE,0.721,
+accuracy,M4 (Hourly),Chronos-T5 (Mid),MASE,0.758,
+accuracy,M4 (Hourly),Chronos-GPT2,MASE,0.710,
+accuracy,M4 (Monthly),Chronos-T5 (Large),MASE,0.960,
+accuracy,M4 (Monthly),Chronos-T5 (Base),MASE,0.970,
+accuracy,M4 (Monthly),Chronos-T5 (Small),MASE,0.982,
+accuracy,M4 (Monthly),Chronos-T5 (Mid),MASE,0.991,
+accuracy,M4 (Monthly),Chronos-GPT2,MASE,1.044,
+accuracy,M4 (Weekly),Chronos-T5 (Large),MASE,1.998,
+accuracy,M4 (Weekly),Chronos-T5 (Base),MASE,2.021,
+accuracy,M4 (Weekly),Chronos-T5 (Small),MASE,2.113,
+accuracy,M4 (Weekly),Chronos-T5 (Mid),MASE,2.152,
+accuracy,M4 (Weekly),Chronos-GPT2,MASE,2.225,
+accuracy,Pedestrian Counts,Chronos-T5 (Large),MASE,0.272,
+accuracy,Pedestrian Counts,Chronos-T5 (Base),MASE,0.286,
+accuracy,Pedestrian Counts,Chronos-T5 (Small),MASE,0.304,
+accuracy,Pedestrian Counts,Chronos-T5 (Mid),MASE,0.303,
+accuracy,Pedestrian Counts,Chronos-GPT2,MASE,0.271,
+accuracy,Rideshare,Chronos-T5 (Large),MASE,0.965,
+accuracy,Rideshare,Chronos-T5 (Base),MASE,0.962,
+accuracy,Rideshare,Chronos-T5 (Small),MASE,0.854,
+accuracy,Rideshare,Chronos-T5 (Mid),MASE,0.830,
+accuracy,Rideshare,Chronos-GPT2,MASE,0.921,
+accuracy,Taxi (30 Min.),Chronos-T5 (Large),MASE,0.930,
+accuracy,Taxi (30 Min.),Chronos-T5 (Base),MASE,0.849,
+accuracy,Taxi (30 Min.),Chronos-T5 (Small),MASE,0.941,
+accuracy,Taxi (30 Min.),Chronos-T5 (Mid),MASE,0.944,
+accuracy,Taxi (30 Min.),Chronos-GPT2,MASE,1.037,
+accuracy,Temperature-Rain,Chronos-T5 (Large),MASE,0.980,
+accuracy,Temperature-Rain,Chronos-T5 (Base),MASE,0.986,
+accuracy,Temperature-Rain,Chronos-T5 (Small),MASE,1.012,
+accuracy,Temperature-Rain,Chronos-T5 (Mid),MASE,1.029,
+accuracy,Temperature-Rain,Chronos-GPT2,MASE,0.974,
+accuracy,Uber TLC (Daily),Chronos-T5 (Large),MASE,0.821,
+accuracy,Uber TLC (Daily),Chronos-T5 (Base),MASE,0.839,
+accuracy,Uber TLC (Daily),Chronos-T5 (Small),MASE,0.870,
+accuracy,Uber TLC (Daily),Chronos-T5 (Mid),MASE,0.906,
+accuracy,Uber TLC (Daily),Chronos-GPT2,MASE,0.835,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Large),MASE,0.670,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Base),MASE,0.673,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Small),MASE,0.677,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Mid),MASE,0.689,
+accuracy,Uber TLC (Hourly),Chronos-GPT2,MASE,0.706,
+accuracy,Agg. Relative Score,Chronos-T5 (Large),MASE,0.695,
+accuracy,Agg. Relative Score,Chronos-T5 (Base),MASE,0.706,
+accuracy,Agg. Relative Score,Chronos-T5 (Small),MASE,0.727,
+accuracy,Agg. Relative Score,Chronos-T5 (Mid),MASE,0.732,
+accuracy,Agg. Relative Score,Chronos-GPT2,MASE,0.741,
+accuracy,Avg. Rank,Chronos-T5 (Large),MASE,3.353,
+accuracy,Avg. Rank,Chronos-T5 (Base),MASE,4.733,
+accuracy,Avg. Rank,Chronos-T5 (Small),MASE,6.067,
+accuracy,Avg. Rank,Chronos-T5 (Mid),MASE,6.467,
+accuracy,Avg. Rank,Chronos-GPT2,MASE,6.933,
+accuracy,Electricity (15 Min.),Chronos-T5 (Large),WQL,0.077,
+accuracy,Electricity (15 Min.),Chronos-T5 (Base),WQL,0.078,
+accuracy,Electricity (15 Min.),Chronos-T5 (Small),WQL,0.080,
+accuracy,Electricity (15 Min.),Chronos-T5 (Mini),WQL,0.082,
+accuracy,Electricity (Hourly),Chronos-T5 (Large),WQL,0.101,
+accuracy,Electricity (Hourly),Chronos-T5 (Base),WQL,0.114,
+accuracy,Electricity (Hourly),Chronos-T5 (Small),WQL,0.105,
+accuracy,Electricity (Hourly),Chronos-T5 (Mini),WQL,0.089,
+accuracy,Electricity (Weekly),Chronos-T5 (Large),WQL,0.059,
+accuracy,Electricity (Weekly),Chronos-T5 (Base),WQL,0.062,
+accuracy,Electricity (Weekly),Chronos-T5 (Small),WQL,0.073,
+accuracy,Electricity (Weekly),Chronos-T5 (Mini),WQL,0.067,
+accuracy,KDD Cup 2018,Chronos-T5 (Large),WQL,0.272,
+accuracy,KDD Cup 2018,Chronos-T5 (Base),WQL,0.268,
+accuracy,KDD Cup 2018,Chronos-T5 (Small),WQL,0.289,
+accuracy,KDD Cup 2018,Chronos-T5 (Mini),WQL,0.271,
+accuracy,London Smart Meters,Chronos-T5 (Large),WQL,0.423,
+accuracy,London Smart Meters,Chronos-T5 (Base),WQL,0.428,
+accuracy,London Smart Meters,Chronos-T5 (Small),WQL,0.431,
+accuracy,London Smart Meters,Chronos-T5 (Mini),WQL,0.436,
+accuracy,M4 (Daily),Chronos-T5 (Large),WQL,0.022,
+accuracy,M4 (Daily),Chronos-T5 (Base),WQL,0.022,
+accuracy,M4 (Daily),Chronos-T5 (Small),WQL,0.022,
+accuracy,M4 (Daily),Chronos-T5 (Mini),WQL,0.022,
+accuracy,M4 (Hourly),Chronos-T5 (Large),WQL,0.022,
+accuracy,M4 (Hourly),Chronos-T5 (Base),WQL,0.024,
+accuracy,M4 (Hourly),Chronos-T5 (Small),WQL,0.024,
+accuracy,M4 (Hourly),Chronos-T5 (Mini),WQL,0.025,
+accuracy,M4 (Monthly),Chronos-T5 (Large),WQL,0.101,
+accuracy,M4 (Monthly),Chronos-T5 (Base),WQL,0.103,
+accuracy,M4 (Monthly),Chronos-T5 (Small),WQL,0.103,
+accuracy,M4 (Monthly),Chronos-T5 (Mini),WQL,0.103,
+accuracy,M4 (Weekly),Chronos-T5 (Large),WQL,0.037,
+accuracy,M4 (Weekly),Chronos-T5 (Base),WQL,0.037,
+accuracy,M4 (Weekly),Chronos-T5 (Small),WQL,0.040,
+accuracy,M4 (Weekly),Chronos-T5 (Mini),WQL,0.041,
+accuracy,Pedestrian Counts,Chronos-T5 (Large),WQL,0.187,
+accuracy,Pedestrian Counts,Chronos-T5 (Base),WQL,0.204,
+accuracy,Pedestrian Counts,Chronos-T5 (Small),WQL,0.237,
+accuracy,Pedestrian Counts,Chronos-T5 (Mini),WQL,0.236,
+accuracy,Rideshare,Chronos-T5 (Large),WQL,0.140,
+accuracy,Rideshare,Chronos-T5 (Base),WQL,0.137,
+accuracy,Rideshare,Chronos-T5 (Small),WQL,0.140,
+accuracy,Rideshare,Chronos-T5 (Mini),WQL,0.133,
+accuracy,Taxi (30 Min.),Chronos-T5 (Large),WQL,0.268,
+accuracy,Taxi (30 Min.),Chronos-T5 (Base),WQL,0.274,
+accuracy,Taxi (30 Min.),Chronos-T5 (Small),WQL,0.312,
+accuracy,Taxi (30 Min.),Chronos-T5 (Mini),WQL,0.313,
+accuracy,Temperature-Rain,Chronos-T5 (Large),WQL,0.663,
+accuracy,Temperature-Rain,Chronos-T5 (Base),WQL,0.669,
+accuracy,Temperature-Rain,Chronos-T5 (Small),WQL,0.685,
+accuracy,Temperature-Rain,Chronos-T5 (Mini),WQL,0.704,
+accuracy,Uber TLC (Daily),Chronos-T5 (Large),WQL,0.096,
+accuracy,Uber TLC (Daily),Chronos-T5 (Base),WQL,0.097,
+accuracy,Uber TLC (Daily),Chronos-T5 (Small),WQL,0.100,
+accuracy,Uber TLC (Daily),Chronos-T5 (Mini),WQL,0.105,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Large),WQL,0.153,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Base),WQL,0.153,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Small),WQL,0.155,
+accuracy,Uber TLC (Hourly),Chronos-T5 (Mini),WQL,0.161,
+accuracy,Agg. Relative Score,Chronos-T5 (Large),WQL,0.564,
+accuracy,Agg. Relative Score,Chronos-T5 (Base),WQL,0.580,
+accuracy,Agg. Relative Score,Chronos-T5 (Small),WQL,0.603,
+accuracy,Agg. Relative Score,Chronos-T5 (Mini),WQL,0.598,
+accuracy,Avg. Rank,Chronos-T5 (Large),WQL,3.400,
+accuracy,Avg. Rank,Chronos-T5 (Base),WQL,4.667,
+accuracy,Avg. Rank,Chronos-T5 (Small),WQL,6.200,
+accuracy,Avg. Rank,Chronos-T5 (Mini),WQL,6.067,
diff --git a/result/per_paper/2403.07815/accuracy_efficiency_traced.csv b/result/per_paper/2403.07815/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ca73c181537cf7c8dcac0bff641accc614c2bebf
--- /dev/null
+++ b/result/per_paper/2403.07815/accuracy_efficiency_traced.csv
@@ -0,0 +1,415 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Australian Electricity,Chronos-T3 (Large),WQL,1.333,,0,2,1
+accuracy,Australian Electricity,Chronos-T3 (Base),WQL,1.319,,0,2,2
+accuracy,Australian Electricity,Chronos-T3 (Base),WQL,1.399,,0,2,3
+accuracy,Australian Electricity,Chronos-T3 (MHI),WQL,1.114,,0,2,4
+accuracy,Australian Electricity,Chronos-GPT2,WQL,1.310,,0,2,5
+accuracy,Car,Chronos-T3 (Large),WQL,0.906,,0,3,1
+accuracy,Car,Chronos-T3 (Base),WQL,0.899,,0,3,2
+accuracy,Car,Chronos-T3 (Base),WQL,0.887,,0,3,3
+accuracy,Car,Chronos-T3 (MHI),WQL,0.893,,0,3,4
+accuracy,Car,Chronos-GPT2,WQL,0.881,,0,3,5
+accuracy,CIF 2016,Chronos-T3 (Large),WQL,0.986,,0,4,1
+accuracy,CIF 2016,Chronos-T3 (Base),WQL,0.981,,0,4,2
+accuracy,CIF 2016,Chronos-T3 (Base),WQL,0.989,,0,4,3
+accuracy,CIF 2016,Chronos-T3 (MHI),WQL,1.053,,0,4,4
+accuracy,CIF 2016,Chronos-GPT2,WQL,1.046,,0,4,5
+accuracy,Covid Deaths,Chronos-T3 (Large),WQL,42.550,,0,5,1
+accuracy,Covid Deaths,Chronos-T3 (Base),WQL,42.687,,0,5,2
+accuracy,Covid Deaths,Chronos-T3 (Base),WQL,42.670,,0,5,3
+accuracy,Covid Deaths,Chronos-T3 (MHI),WQL,43.621,,0,5,4
+accuracy,Covid Deaths,Chronos-GPT2,WQL,48.215,,0,5,5
+accuracy,Dominick,Chronos-T3 (Large),WQL,0.818,,0,6,1
+accuracy,Dominick,Chronos-T3 (Base),WQL,0.816,,0,6,2
+accuracy,Dominick,Chronos-T3 (Base),WQL,0.819,,0,6,3
+accuracy,Dominick,Chronos-T3 (MHI),WQL,0.833,,0,6,4
+accuracy,Dominick,Chronos-GPT2,WQL,0.820,,0,6,5
+accuracy,ERCOT Load,Chronos-T3 (Large),WQL,0.617,,0,7,1
+accuracy,ERCOT Load,Chronos-T3 (Base),WQL,0.550,,0,7,2
+accuracy,ERCOT Load,Chronos-T3 (Base),WQL,0.573,,0,7,3
+accuracy,ERCOT Load,Chronos-T3 (MHI),WQL,0.588,,0,7,4
+accuracy,ERCOT Load,Chronos-GPT2,WQL,0.561,,0,7,5
+accuracy,ETT (15 Min.),Chronos-T3 (Large),WQL,0.741,,0,8,1
+accuracy,ETT (15 Min.),Chronos-T3 (Base),WQL,0.739,,0,8,2
+accuracy,ETT (15 Min.),Chronos-T3 (Base),WQL,0.710,,0,8,3
+accuracy,ETT (15 Min.),Chronos-T3 (MHI),WQL,0.790,,0,8,4
+accuracy,ETT (15 Min.),Chronos-GPT2,WQL,0.796,,0,8,5
+accuracy,ETT (Hourly),Chronos-T3 (Large),WQL,0.735,,0,9,1
+accuracy,ETT (Hourly),Chronos-T3 (Base),WQL,0.789,,0,9,2
+accuracy,ETT (Hourly),Chronos-T3 (Base),WQL,0.789,,0,9,3
+accuracy,ETT (Hourly),Chronos-T3 (MHI),WQL,0.797,,0,9,4
+accuracy,ETT (Hourly),Chronos-GPT2,WQL,0.768,,0,9,5
+accuracy,Exchange Rate,Chronos-T3 (Large),WQL,2.975,,0,10,1
+accuracy,Exchange Rate,Chronos-T3 (Base),WQL,2.433,,0,10,2
+accuracy,Exchange Rate,Chronos-T3 (Base),WQL,2.252,,0,10,3
+accuracy,Exchange Rate,Chronos-T3 (MHI),WQL,2.030,,0,10,4
+accuracy,Exchange Rate,Chronos-GPT2,WQL,2.335,,0,10,5
+accuracy,FRED-MD,Chronos-T3 (Large),WQL,0.500,,0,11,1
+accuracy,FRED-MD,Chronos-T3 (Base),WQL,0.486,,0,11,2
+accuracy,FRED-MD,Chronos-T3 (Base),WQL,0.496,,0,11,3
+accuracy,FRED-MD,Chronos-T3 (MHI),WQL,0.483,,0,11,4
+accuracy,FRED-MD,Chronos-GPT2,WQL,0.468,,0,11,5
+accuracy,Hospital,Chronos-T3 (Large),WQL,0.810,,0,12,1
+accuracy,Hospital,Chronos-T3 (Base),WQL,0.810,,0,12,2
+accuracy,Hospital,Chronos-T3 (Base),WQL,0.815,,0,12,3
+accuracy,Hospital,Chronos-T3 (MHI),WQL,0.817,,0,12,4
+accuracy,Hospital,Chronos-GPT2,WQL,0.831,,0,12,5
+accuracy,MI (Monthly),Chronos-T3 (Large),WQL,1.090,,0,13,1
+accuracy,MI (Monthly),Chronos-T3 (Base),WQL,1.117,,0,13,2
+accuracy,MI (Monthly),Chronos-T3 (Base),WQL,1.169,,0,13,3
+accuracy,MI (Monthly),Chronos-T3 (MHI),WQL,1.174,,0,13,4
+accuracy,MI (Monthly),Chronos-GPT2,WQL,1.182,,0,13,5
+accuracy,MI (Quarterly),Chronos-T3 (Large),WQL,1.713,,0,14,1
+accuracy,MI (Quarterly),Chronos-T3 (Base),WQL,1.739,,0,14,2
+accuracy,MI (Quarterly),Chronos-T3 (Base),WQL,1.764,,0,14,3
+accuracy,MI (Quarterly),Chronos-T3 (MHI),WQL,1.785,,0,14,4
+accuracy,MI (Quarterly),Chronos-GPT2,WQL,1.785,,0,14,5
+accuracy,MI (Yearly),Chronos-T3 (Large),WQL,4.301,,0,15,1
+accuracy,MI (Yearly),Chronos-T3 (Base),WQL,4.624,,0,15,2
+accuracy,MI (Yearly),Chronos-T3 (Base),WQL,4.659,,0,15,3
+accuracy,MI (Yearly),Chronos-T3 (MHI),WQL,4.956,,0,15,4
+accuracy,MI (Yearly),Chronos-GPT2,WQL,4.751,,0,15,5
+accuracy,M3 (Monthly),Chronos-T3 (Large),WQL,0.857,,0,16,1
+accuracy,M3 (Monthly),Chronos-T3 (Base),WQL,0.868,,0,16,2
+accuracy,M3 (Monthly),Chronos-T3 (Base),WQL,0.885,,0,16,3
+accuracy,M3 (Monthly),Chronos-T3 (MHI),WQL,0.900,,0,16,4
+accuracy,M3 (Monthly),Chronos-GPT2,WQL,0.930,,0,16,5
+accuracy,M3 (Quarterly),Chronos-T3 (Large),WQL,1.181,,0,17,1
+accuracy,M3 (Quarterly),Chronos-T3 (Base),WQL,1.199,,0,17,2
+accuracy,M3 (Quarterly),Chronos-T3 (Base),WQL,1.256,,0,17,3
+accuracy,M3 (Quarterly),Chronos-T3 (MHI),WQL,1.289,,0,17,4
+accuracy,M3 (Quarterly),Chronos-GPT2,WQL,1.241,,0,17,5
+accuracy,M3 (Yearly),Chronos-T3 (Large),WQL,3.106,,0,18,1
+accuracy,M3 (Yearly),Chronos-T3 (Base),WQL,3.269,,0,18,2
+accuracy,M3 (Yearly),Chronos-T3 (Base),WQL,3.276,,0,18,3
+accuracy,M3 (Yearly),Chronos-T3 (MHI),WQL,3.305,,0,18,4
+accuracy,M3 (Yearly),Chronos-GPT2,WQL,3.158,,0,18,5
+accuracy,M4 (Quarterly),Chronos-T3 (Large),WQL,1.216,,0,19,1
+accuracy,M4 (Quarterly),Chronos-T3 (Base),WQL,1.231,,0,19,2
+accuracy,M4 (Quarterly),Chronos-T3 (Base),WQL,1.246,,0,19,3
+accuracy,M4 (Quarterly),Chronos-T3 (MHI),WQL,1.271,,0,19,4
+accuracy,M4 (Quarterly),Chronos-GPT2,WQL,1.312,,0,19,5
+accuracy,M4 (Yearly),Chronos-T3 (Large),WQL,3.606,,0,20,1
+accuracy,M4 (Yearly),Chronos-T3 (Base),WQL,3.678,,0,20,2
+accuracy,M4 (Yearly),Chronos-T3 (Base),WQL,3.651,,0,20,3
+accuracy,M4 (Yearly),Chronos-T3 (MHI),WQL,3.743,,0,20,4
+accuracy,M4 (Yearly),Chronos-GPT2,WQL,3.933,,0,20,5
+accuracy,M5,Chronos-T3 (Large),WQL,0.944,,0,21,1
+accuracy,M5,Chronos-T3 (Base),WQL,0.939,,0,21,2
+accuracy,M5,Chronos-T3 (Base),WQL,0.940,,0,21,3
+accuracy,M5,Chronos-T3 (MHI),WQL,0.944,,0,21,4
+accuracy,M5,Chronos-GPT2,WQL,0.969,,0,21,5
+accuracy,NNS (Daily),Chronos-T3 (Large),WQL,0.573,,0,22,1
+accuracy,NNS (Daily),Chronos-T3 (Base),WQL,0.585,,0,22,2
+accuracy,NNS (Daily),Chronos-T3 (Base),WQL,0.615,,0,22,3
+accuracy,NNS (Daily),Chronos-T3 (MHI),WQL,0.642,,0,22,4
+accuracy,NNS (Daily),Chronos-GPT2,WQL,0.661,,0,22,5
+accuracy,NNS (Weekly),Chronos-T3 (Large),WQL,0.940,,0,23,1
+accuracy,NNS (Weekly),Chronos-T3 (Base),WQL,0.938,,0,23,2
+accuracy,NNS (Weekly),Chronos-T3 (Base),WQL,0.944,,0,23,3
+accuracy,NNS (Weekly),Chronos-T3 (MHI),WQL,0.947,,0,23,4
+accuracy,NNS (Weekly),Chronos-GPT2,WQL,0.963,,0,23,5
+accuracy,Tourism (Monthly),Chronos-T3 (Large),WQL,1.761,,0,24,1
+accuracy,Tourism (Monthly),Chronos-T3 (Base),WQL,1.828,,0,24,2
+accuracy,Tourism (Monthly),Chronos-T3 (Base),WQL,1.900,,0,24,3
+accuracy,Tourism (Monthly),Chronos-T3 (MHI),WQL,1.950,,0,24,4
+accuracy,Tourism (Monthly),Chronos-GPT2,WQL,1.783,,0,24,5
+accuracy,Tourism (Quarterly),Chronos-T3 (Large),WQL,1.677,,0,25,1
+accuracy,Tourism (Quarterly),Chronos-T3 (Base),WQL,1.717,,0,25,2
+accuracy,Tourism (Quarterly),Chronos-T3 (Base),WQL,1.730,,0,25,3
+accuracy,Tourism (Quarterly),Chronos-T3 (MHI),WQL,1.829,,0,25,4
+accuracy,Tourism (Quarterly),Chronos-GPT2,WQL,1.828,,0,25,5
+accuracy,Tourism (Yearly),Chronos-T3 (Large),WQL,3.755,,0,26,1
+accuracy,Tourism (Yearly),Chronos-T3 (Base),WQL,3.900,,0,26,2
+accuracy,Tourism (Yearly),Chronos-T3 (Base),WQL,3.901,,0,26,3
+accuracy,Tourism (Yearly),Chronos-T3 (MHI),WQL,4.048,,0,26,4
+accuracy,Tourism (Yearly),Chronos-GPT2,WQL,3.882,,0,26,5
+accuracy,Traffic,Chronos-T3 (Large),WQL,0.804,,0,27,1
+accuracy,Traffic,Chronos-T3 (Base),WQL,0.828,,0,27,2
+accuracy,Traffic,Chronos-T3 (Base),WQL,0.837,,0,27,3
+accuracy,Traffic,Chronos-T3 (MHI),WQL,0.850,,0,27,4
+accuracy,Traffic,Chronos-GPT2,WQL,0.818,,0,27,5
+accuracy,Weather,Chronos-T3 (Large),WQL,0.822,,0,28,1
+accuracy,Weather,Chronos-T3 (Base),WQL,0.824,,0,28,2
+accuracy,Weather,Chronos-T3 (Base),WQL,0.836,,0,28,3
+accuracy,Weather,Chronos-T3 (MHI),WQL,0.853,,0,28,4
+accuracy,Weather,Chronos-GPT2,WQL,0.858,,0,28,5
+accuracy,Avg. Relative Score,Chronos-T3 (Large),WQL,0.823,,0,29,1
+accuracy,Avg. Relative Score,Chronos-T3 (Base),WQL,0.832,,0,29,2
+accuracy,Avg. Relative Score,Chronos-T3 (Base),WQL,0.841,,0,29,3
+accuracy,Avg. Relative Score,Chronos-T3 (MHI),WQL,0.850,,0,29,4
+accuracy,Avg. Relative Score,Chronos-GPT2,WQL,0.852,,0,29,5
+accuracy,Avg. Rank,Chronos-T3 (Large),WQL,5.351,,0,30,1
+accuracy,Avg. Rank,Chronos-T3 (Base),WQL,9.296,,0,30,2
+accuracy,Avg. Rank,Chronos-T3 (Base),WQL,10.593,,0,30,3
+accuracy,Avg. Rank,Chronos-T3 (MHI),WQL,12.037,,0,30,4
+accuracy,Avg. Rank,Chronos-GPT2,WQL,11.630,,0,30,5
+accuracy,Australian Electricity,Chronos-T5(Large),WQL,0.067,,1,2,1
+accuracy,Australian Electricity,Chronos-T5(Basic),WQL,0.075,,1,2,2
+accuracy,Australian Electricity,Chronos-T5(Small),WQL,0.074,,1,2,3
+accuracy,Australian Electricity,Chronos-T5(Mini),WQL,0.063,,1,2,4
+accuracy,Car Parts,Chronos-T5(Large),WQL,1.060,,1,3,1
+accuracy,Car Parts,Chronos-T5(Basic),WQL,1.057,,1,3,2
+accuracy,Car Parts,Chronos-T5(Small),WQL,1.029,,1,3,3
+accuracy,Car Parts,Chronos-T5(Mini),WQL,1.024,,1,3,4
+accuracy,CIF 2016,Chronos-T5(Large),WQL,0.014,,1,4,1
+accuracy,CIF 2016,Chronos-T5(Basic),WQL,0.013,,1,4,2
+accuracy,CIF 2016,Chronos-T5(Small),WQL,0.015,,1,4,3
+accuracy,CIF 2016,Chronos-T5(Mini),WQL,0.013,,1,4,4
+accuracy,Covid Deaths,Chronos-T5(Large),WQL,0.045,,1,5,1
+accuracy,Covid Deaths,Chronos-T5(Basic),WQL,0.048,,1,5,2
+accuracy,Covid Deaths,Chronos-T5(Small),WQL,0.059,,1,5,3
+accuracy,Covid Deaths,Chronos-T5(Mini),WQL,0.084,,1,5,4
+accuracy,Donnick,Chronos-T5(Large),WQL,0.332,,1,6,1
+accuracy,Donnick,Chronos-T5(Basic),WQL,0.333,,1,6,2
+accuracy,Donnick,Chronos-T5(Small),WQL,0.338,,1,6,3
+accuracy,Donnick,Chronos-T5(Mini),WQL,0.346,,1,6,4
+accuracy,ERCOT Load,Chronos-T5(Large),WQL,0.019,,1,7,1
+accuracy,ERCOT Load,Chronos-T5(Basic),WQL,0.016,,1,7,2
+accuracy,ERCOT Load,Chronos-T5(Small),WQL,0.018,,1,7,3
+accuracy,ERCOT Load,Chronos-T5(Mini),WQL,0.018,,1,7,4
+accuracy,ETT (15 Min.),Chronos-T5(Large),WQL,0.068,,1,8,1
+accuracy,ETT (15 Min.),Chronos-T5(Basic),WQL,0.069,,1,8,2
+accuracy,ETT (15 Min.),Chronos-T5(Small),WQL,0.064,,1,8,3
+accuracy,ETT (15 Min.),Chronos-T5(Mini),WQL,0.072,,1,8,4
+accuracy,ETT (Hourly),Chronos-T5(Large),WQL,0.073,,1,9,1
+accuracy,ETT (Hourly),Chronos-T5(Basic),WQL,0.081,,1,9,2
+accuracy,ETT (Hourly),Chronos-T5(Small),WQL,0.080,,1,9,3
+accuracy,ETT (Hourly),Chronos-T5(Mini),WQL,0.085,,1,9,4
+accuracy,Exchange Rate,Chronos-T5(Large),WQL,0.013,,1,10,1
+accuracy,Exchange Rate,Chronos-T5(Basic),WQL,0.014,,1,10,2
+accuracy,Exchange Rate,Chronos-T5(Small),WQL,0.013,,1,10,3
+accuracy,Exchange Rate,Chronos-T5(Mini),WQL,0.012,,1,10,4
+accuracy,FRED-MD,Chronos-T5(Large),WQL,0.020,,1,11,1
+accuracy,FRED-MD,Chronos-T5(Basic),WQL,0.022,,1,11,2
+accuracy,FRED-MD,Chronos-T5(Small),WQL,0.017,,1,11,3
+accuracy,FRED-MD,Chronos-T5(Mini),WQL,0.017,,1,11,4
+accuracy,Hospital,Chronos-T5(Large),WQL,0.056,,1,12,1
+accuracy,Hospital,Chronos-T5(Basic),WQL,0.056,,1,12,2
+accuracy,Hospital,Chronos-T5(Small),WQL,0.057,,1,12,3
+accuracy,Hospital,Chronos-T5(Mini),WQL,0.058,,1,12,4
+accuracy,M1 (Monthly),Chronos-T5(Large),WQL,0.130,,1,13,1
+accuracy,M1 (Monthly),Chronos-T5(Basic),WQL,0.128,,1,13,2
+accuracy,M1 (Monthly),Chronos-T5(Small),WQL,0.139,,1,13,3
+accuracy,M1 (Monthly),Chronos-T5(Mini),WQL,0.138,,1,13,4
+accuracy,M1 (Quarterly),Chronos-T5(Large),WQL,0.107,,1,14,1
+accuracy,M1 (Quarterly),Chronos-T5(Basic),WQL,0.105,,1,14,2
+accuracy,M1 (Quarterly),Chronos-T5(Small),WQL,0.103,,1,14,3
+accuracy,M1 (Quarterly),Chronos-T5(Mini),WQL,0.103,,1,14,4
+accuracy,M1 (Yearly),Chronos-T5(Large),WQL,0.183,,1,15,1
+accuracy,M1 (Yearly),Chronos-T5(Basic),WQL,0.181,,1,15,2
+accuracy,M1 (Yearly),Chronos-T5(Small),WQL,0.172,,1,15,3
+accuracy,M1 (Yearly),Chronos-T5(Mini),WQL,0.179,,1,15,4
+accuracy,M3 (Monthly),Chronos-T5(Large),WQL,0.096,,1,16,1
+accuracy,M3 (Monthly),Chronos-T5(Basic),WQL,0.097,,1,16,2
+accuracy,M3 (Monthly),Chronos-T5(Small),WQL,0.100,,1,16,3
+accuracy,M3 (Monthly),Chronos-T5(Mini),WQL,0.099,,1,16,4
+accuracy,M3 (Quarterly),Chronos-T5(Large),WQL,0.074,,1,17,1
+accuracy,M3 (Quarterly),Chronos-T5(Basic),WQL,0.076,,1,17,2
+accuracy,M3 (Quarterly),Chronos-T5(Small),WQL,0.079,,1,17,3
+accuracy,M3 (Quarterly),Chronos-T5(Mini),WQL,0.081,,1,17,4
+accuracy,M3 (Yearly),Chronos-T5(Large),WQL,0.151,,1,18,1
+accuracy,M3 (Yearly),Chronos-T5(Basic),WQL,0.153,,1,18,2
+accuracy,M3 (Yearly),Chronos-T5(Small),WQL,0.155,,1,18,3
+accuracy,M3 (Yearly),Chronos-T5(Mini),WQL,0.159,,1,18,4
+accuracy,M4 (Quarterly),Chronos-T5(Large),WQL,0.082,,1,19,1
+accuracy,M4 (Quarterly),Chronos-T5(Basic),WQL,0.083,,1,19,2
+accuracy,M4 (Quarterly),Chronos-T5(Small),WQL,0.084,,1,19,3
+accuracy,M4 (Quarterly),Chronos-T5(Mini),WQL,0.086,,1,19,4
+accuracy,M4 (Yearly),Chronos-T5(Large),WQL,0.134,,1,20,1
+accuracy,M4 (Yearly),Chronos-T5(Basic),WQL,0.137,,1,20,2
+accuracy,M4 (Yearly),Chronos-T5(Small),WQL,0.136,,1,20,3
+accuracy,M4 (Yearly),Chronos-T5(Mini),WQL,0.140,,1,20,4
+accuracy,M5,Chronos-T5(Large),WQL,0.587,,1,21,1
+accuracy,M5,Chronos-T5(Basic),WQL,0.586,,1,21,2
+accuracy,M5,Chronos-T5(Small),WQL,0.590,,1,21,3
+accuracy,M5,Chronos-T5(Mini),WQL,0.595,,1,21,4
+accuracy,NN5 (Daily),Chronos-T5(Large),WQL,0.156,,1,22,1
+accuracy,NN5 (Daily),Chronos-T5(Basic),WQL,0.161,,1,22,2
+accuracy,NN5 (Daily),Chronos-T5(Small),WQL,0.169,,1,22,3
+accuracy,NN5 (Daily),Chronos-T5(Mini),WQL,0.173,,1,22,4
+accuracy,NN5 (Weekly),Chronos-T5(Large),WQL,0.091,,1,23,1
+accuracy,NN5 (Weekly),Chronos-T5(Basic),WQL,0.091,,1,23,2
+accuracy,NN5 (Weekly),Chronos-T5(Small),WQL,0.090,,1,23,3
+accuracy,NN5 (Weekly),Chronos-T5(Mini),WQL,0.091,,1,23,4
+accuracy,Tourism (Monthly),Chronos-T5(Large),WQL,0.100,,1,24,1
+accuracy,Tourism (Monthly),Chronos-T5(Basic),WQL,0.103,,1,24,2
+accuracy,Tourism (Monthly),Chronos-T5(Small),WQL,0.113,,1,24,3
+accuracy,Tourism (Monthly),Chronos-T5(Mini),WQL,0.109,,1,24,4
+accuracy,Tourism (Quarterly),Chronos-T5(Large),WQL,0.061,,1,25,1
+accuracy,Tourism (Quarterly),Chronos-T5(Basic),WQL,0.069,,1,25,2
+accuracy,Tourism (Quarterly),Chronos-T5(Small),WQL,0.069,,1,25,3
+accuracy,Tourism (Quarterly),Chronos-T5(Mini),WQL,0.074,,1,25,4
+accuracy,Tourism (Yearly),Chronos-T5(Large),WQL,0.183,,1,26,1
+accuracy,Tourism (Yearly),Chronos-T5(Basic),WQL,0.207,,1,26,2
+accuracy,Tourism (Yearly),Chronos-T5(Small),WQL,0.200,,1,26,3
+accuracy,Tourism (Yearly),Chronos-T5(Mini),WQL,0.218,,1,26,4
+accuracy,Traffic,Chronos-T5(Large),WQL,0.256,,1,27,1
+accuracy,Traffic,Chronos-T5(Basic),WQL,0.264,,1,27,2
+accuracy,Traffic,Chronos-T5(Small),WQL,0.263,,1,27,3
+accuracy,Traffic,Chronos-T5(Mini),WQL,0.264,,1,27,4
+accuracy,Weather,Chronos-T5(Large),WQL,0.139,,1,28,1
+accuracy,Weather,Chronos-T5(Basic),WQL,0.140,,1,28,2
+accuracy,Weather,Chronos-T5(Small),WQL,0.143,,1,28,3
+accuracy,Weather,Chronos-T5(Mini),WQL,0.150,,1,28,4
+accuracy,Agg. Relative Score,Chronos-T5(Large),WQL,0.645,,1,29,1
+accuracy,Agg. Relative Score,Chronos-T5(Basic),WQL,0.662,,1,29,2
+accuracy,Agg. Relative Score,Chronos-T5(Small),WQL,0.667,,1,29,3
+accuracy,Agg. Relative Score,Chronos-T5(Mini),WQL,0.678,,1,29,4
+accuracy,Avg. Rank,Chronos-T5(Large),WQL,5.333,,1,30,1
+accuracy,Avg. Rank,Chronos-T5(Basic),WQL,9.407,,1,30,2
+accuracy,Avg. Rank,Chronos-T5(Small),WQL,9.889,,1,30,3
+accuracy,Avg. Rank,Chronos-T5(Mini),WQL,11.296,,1,30,4
+accuracy,Electricity (15 Min.),Chronos-T5 (Large),MASE,0.391,,2,2,1
+accuracy,Electricity (15 Min.),Chronos-T5 (Base),MASE,0.394,,2,2,2
+accuracy,Electricity (15 Min.),Chronos-T5 (Small),MASE,0.418,,2,2,3
+accuracy,Electricity (15 Min.),Chronos-T5 (Mid),MASE,0.445,,2,2,4
+accuracy,Electricity (15 Min.),Chronos-GPT2,MASE,0.388,,2,2,5
+accuracy,Electricity (Hourly),Chronos-T5 (Large),MASE,1.439,,2,3,1
+accuracy,Electricity (Hourly),Chronos-T5 (Base),MASE,1.590,,2,3,2
+accuracy,Electricity (Hourly),Chronos-T5 (Small),MASE,1.477,,2,3,3
+accuracy,Electricity (Hourly),Chronos-T5 (Mid),MASE,1.348,,2,3,4
+accuracy,Electricity (Hourly),Chronos-GPT2,MASE,1.636,,2,3,5
+accuracy,Electricity (Weekly),Chronos-T5 (Large),MASE,1.739,,2,4,1
+accuracy,Electricity (Weekly),Chronos-T5 (Base),MASE,1.801,,2,4,2
+accuracy,Electricity (Weekly),Chronos-T5 (Small),MASE,1.942,,2,4,3
+accuracy,Electricity (Weekly),Chronos-T5 (Mid),MASE,1.954,,2,4,4
+accuracy,Electricity (Weekly),Chronos-GPT2,MASE,1.720,,2,4,5
+accuracy,KDD Cap 2018,Chronos-T5 (Large),MASE,0.683,,2,5,1
+accuracy,KDD Cap 2018,Chronos-T5 (Base),MASE,0.646,,2,5,2
+accuracy,KDD Cap 2018,Chronos-T5 (Small),MASE,0.687,,2,5,3
+accuracy,KDD Cap 2018,Chronos-T5 (Mid),MASE,0.667,,2,5,4
+accuracy,KDD Cap 2018,Chronos-GPT2,MASE,0.881,,2,5,5
+accuracy,London Smart Meters,Chronos-T5 (Large),MASE,0.828,,2,6,1
+accuracy,London Smart Meters,Chronos-T5 (Base),MASE,0.838,,2,6,2
+accuracy,London Smart Meters,Chronos-T5 (Small),MASE,0.846,,2,6,3
+accuracy,London Smart Meters,Chronos-T5 (Mid),MASE,0.857,,2,6,4
+accuracy,London Smart Meters,Chronos-GPT2,MASE,0.842,,2,6,5
+accuracy,M4 (Daily),Chronos-T5 (Large),MASE,3.144,,2,7,1
+accuracy,M4 (Daily),Chronos-T5 (Base),MASE,3.160,,2,7,2
+accuracy,M4 (Daily),Chronos-T5 (Small),MASE,3.148,,2,7,3
+accuracy,M4 (Daily),Chronos-T5 (Mid),MASE,3.154,,2,7,4
+accuracy,M4 (Daily),Chronos-GPT2,MASE,3.079,,2,7,5
+accuracy,M4 (Hourly),Chronos-T5 (Large),MASE,0.682,,2,8,1
+accuracy,M4 (Hourly),Chronos-T5 (Base),MASE,0.694,,2,8,2
+accuracy,M4 (Hourly),Chronos-T5 (Small),MASE,0.721,,2,8,3
+accuracy,M4 (Hourly),Chronos-T5 (Mid),MASE,0.758,,2,8,4
+accuracy,M4 (Hourly),Chronos-GPT2,MASE,0.710,,2,8,5
+accuracy,M4 (Monthly),Chronos-T5 (Large),MASE,0.960,,2,9,1
+accuracy,M4 (Monthly),Chronos-T5 (Base),MASE,0.970,,2,9,2
+accuracy,M4 (Monthly),Chronos-T5 (Small),MASE,0.982,,2,9,3
+accuracy,M4 (Monthly),Chronos-T5 (Mid),MASE,0.991,,2,9,4
+accuracy,M4 (Monthly),Chronos-GPT2,MASE,1.044,,2,9,5
+accuracy,M4 (Weekly),Chronos-T5 (Large),MASE,1.998,,2,10,1
+accuracy,M4 (Weekly),Chronos-T5 (Base),MASE,2.021,,2,10,2
+accuracy,M4 (Weekly),Chronos-T5 (Small),MASE,2.113,,2,10,3
+accuracy,M4 (Weekly),Chronos-T5 (Mid),MASE,2.152,,2,10,4
+accuracy,M4 (Weekly),Chronos-GPT2,MASE,2.225,,2,10,5
+accuracy,Pedestrian Counts,Chronos-T5 (Large),MASE,0.272,,2,11,1
+accuracy,Pedestrian Counts,Chronos-T5 (Base),MASE,0.286,,2,11,2
+accuracy,Pedestrian Counts,Chronos-T5 (Small),MASE,0.304,,2,11,3
+accuracy,Pedestrian Counts,Chronos-T5 (Mid),MASE,0.303,,2,11,4
+accuracy,Pedestrian Counts,Chronos-GPT2,MASE,0.271,,2,11,5
+accuracy,Rideshare,Chronos-T5 (Large),MASE,0.965,,2,12,1
+accuracy,Rideshare,Chronos-T5 (Base),MASE,0.962,,2,12,2
+accuracy,Rideshare,Chronos-T5 (Small),MASE,0.854,,2,12,3
+accuracy,Rideshare,Chronos-T5 (Mid),MASE,0.830,,2,12,4
+accuracy,Rideshare,Chronos-GPT2,MASE,0.921,,2,12,5
+accuracy,Taxi (30 Min.),Chronos-T5 (Large),MASE,0.930,,2,13,1
+accuracy,Taxi (30 Min.),Chronos-T5 (Base),MASE,0.849,,2,13,2
+accuracy,Taxi (30 Min.),Chronos-T5 (Small),MASE,0.941,,2,13,3
+accuracy,Taxi (30 Min.),Chronos-T5 (Mid),MASE,0.944,,2,13,4
+accuracy,Taxi (30 Min.),Chronos-GPT2,MASE,1.037,,2,13,5
+accuracy,Temperature-Rain,Chronos-T5 (Large),MASE,0.980,,2,14,1
+accuracy,Temperature-Rain,Chronos-T5 (Base),MASE,0.986,,2,14,2
+accuracy,Temperature-Rain,Chronos-T5 (Small),MASE,1.012,,2,14,3
+accuracy,Temperature-Rain,Chronos-T5 (Mid),MASE,1.029,,2,14,4
+accuracy,Temperature-Rain,Chronos-GPT2,MASE,0.974,,2,14,5
+accuracy,Uber TLC (Daily),Chronos-T5 (Large),MASE,0.821,,2,15,1
+accuracy,Uber TLC (Daily),Chronos-T5 (Base),MASE,0.839,,2,15,2
+accuracy,Uber TLC (Daily),Chronos-T5 (Small),MASE,0.870,,2,15,3
+accuracy,Uber TLC (Daily),Chronos-T5 (Mid),MASE,0.906,,2,15,4
+accuracy,Uber TLC (Daily),Chronos-GPT2,MASE,0.835,,2,15,5
+accuracy,Uber TLC (Hourly),Chronos-T5 (Large),MASE,0.670,,2,16,1
+accuracy,Uber TLC (Hourly),Chronos-T5 (Base),MASE,0.673,,2,16,2
+accuracy,Uber TLC (Hourly),Chronos-T5 (Small),MASE,0.677,,2,16,3
+accuracy,Uber TLC (Hourly),Chronos-T5 (Mid),MASE,0.689,,2,16,4
+accuracy,Uber TLC (Hourly),Chronos-GPT2,MASE,0.706,,2,16,5
+accuracy,Agg. Relative Score,Chronos-T5 (Large),MASE,0.695,,2,17,1
+accuracy,Agg. Relative Score,Chronos-T5 (Base),MASE,0.706,,2,17,2
+accuracy,Agg. Relative Score,Chronos-T5 (Small),MASE,0.727,,2,17,3
+accuracy,Agg. Relative Score,Chronos-T5 (Mid),MASE,0.732,,2,17,4
+accuracy,Agg. Relative Score,Chronos-GPT2,MASE,0.741,,2,17,5
+accuracy,Avg. Rank,Chronos-T5 (Large),MASE,3.353,,2,18,1
+accuracy,Avg. Rank,Chronos-T5 (Base),MASE,4.733,,2,18,2
+accuracy,Avg. Rank,Chronos-T5 (Small),MASE,6.067,,2,18,3
+accuracy,Avg. Rank,Chronos-T5 (Mid),MASE,6.467,,2,18,4
+accuracy,Avg. Rank,Chronos-GPT2,MASE,6.933,,2,18,5
+accuracy,Electricity (15 Min.),Chronos-T5 (Large),WQL,0.077,,3,2,1
+accuracy,Electricity (15 Min.),Chronos-T5 (Base),WQL,0.078,,3,2,2
+accuracy,Electricity (15 Min.),Chronos-T5 (Small),WQL,0.080,,3,2,3
+accuracy,Electricity (15 Min.),Chronos-T5 (Mini),WQL,0.082,,3,2,4
+accuracy,Electricity (Hourly),Chronos-T5 (Large),WQL,0.101,,3,3,1
+accuracy,Electricity (Hourly),Chronos-T5 (Base),WQL,0.114,,3,3,2
+accuracy,Electricity (Hourly),Chronos-T5 (Small),WQL,0.105,,3,3,3
+accuracy,Electricity (Hourly),Chronos-T5 (Mini),WQL,0.089,,3,3,4
+accuracy,Electricity (Weekly),Chronos-T5 (Large),WQL,0.059,,3,4,1
+accuracy,Electricity (Weekly),Chronos-T5 (Base),WQL,0.062,,3,4,2
+accuracy,Electricity (Weekly),Chronos-T5 (Small),WQL,0.073,,3,4,3
+accuracy,Electricity (Weekly),Chronos-T5 (Mini),WQL,0.067,,3,4,4
+accuracy,KDD Cup 2018,Chronos-T5 (Large),WQL,0.272,,3,5,1
+accuracy,KDD Cup 2018,Chronos-T5 (Base),WQL,0.268,,3,5,2
+accuracy,KDD Cup 2018,Chronos-T5 (Small),WQL,0.289,,3,5,3
+accuracy,KDD Cup 2018,Chronos-T5 (Mini),WQL,0.271,,3,5,4
+accuracy,London Smart Meters,Chronos-T5 (Large),WQL,0.423,,3,6,1
+accuracy,London Smart Meters,Chronos-T5 (Base),WQL,0.428,,3,6,2
+accuracy,London Smart Meters,Chronos-T5 (Small),WQL,0.431,,3,6,3
+accuracy,London Smart Meters,Chronos-T5 (Mini),WQL,0.436,,3,6,4
+accuracy,M4 (Daily),Chronos-T5 (Large),WQL,0.022,,3,7,1
+accuracy,M4 (Daily),Chronos-T5 (Base),WQL,0.022,,3,7,2
+accuracy,M4 (Daily),Chronos-T5 (Small),WQL,0.022,,3,7,3
+accuracy,M4 (Daily),Chronos-T5 (Mini),WQL,0.022,,3,7,4
+accuracy,M4 (Hourly),Chronos-T5 (Large),WQL,0.022,,3,8,1
+accuracy,M4 (Hourly),Chronos-T5 (Base),WQL,0.024,,3,8,2
+accuracy,M4 (Hourly),Chronos-T5 (Small),WQL,0.024,,3,8,3
+accuracy,M4 (Hourly),Chronos-T5 (Mini),WQL,0.025,,3,8,4
+accuracy,M4 (Monthly),Chronos-T5 (Large),WQL,0.101,,3,9,1
+accuracy,M4 (Monthly),Chronos-T5 (Base),WQL,0.103,,3,9,2
+accuracy,M4 (Monthly),Chronos-T5 (Small),WQL,0.103,,3,9,3
+accuracy,M4 (Monthly),Chronos-T5 (Mini),WQL,0.103,,3,9,4
+accuracy,M4 (Weekly),Chronos-T5 (Large),WQL,0.037,,3,10,1
+accuracy,M4 (Weekly),Chronos-T5 (Base),WQL,0.037,,3,10,2
+accuracy,M4 (Weekly),Chronos-T5 (Small),WQL,0.040,,3,10,3
+accuracy,M4 (Weekly),Chronos-T5 (Mini),WQL,0.041,,3,10,4
+accuracy,Pedestrian Counts,Chronos-T5 (Large),WQL,0.187,,3,11,1
+accuracy,Pedestrian Counts,Chronos-T5 (Base),WQL,0.204,,3,11,2
+accuracy,Pedestrian Counts,Chronos-T5 (Small),WQL,0.237,,3,11,3
+accuracy,Pedestrian Counts,Chronos-T5 (Mini),WQL,0.236,,3,11,4
+accuracy,Rideshare,Chronos-T5 (Large),WQL,0.140,,3,12,1
+accuracy,Rideshare,Chronos-T5 (Base),WQL,0.137,,3,12,2
+accuracy,Rideshare,Chronos-T5 (Small),WQL,0.140,,3,12,3
+accuracy,Rideshare,Chronos-T5 (Mini),WQL,0.133,,3,12,4
+accuracy,Taxi (30 Min.),Chronos-T5 (Large),WQL,0.268,,3,13,1
+accuracy,Taxi (30 Min.),Chronos-T5 (Base),WQL,0.274,,3,13,2
+accuracy,Taxi (30 Min.),Chronos-T5 (Small),WQL,0.312,,3,13,3
+accuracy,Taxi (30 Min.),Chronos-T5 (Mini),WQL,0.313,,3,13,4
+accuracy,Temperature-Rain,Chronos-T5 (Large),WQL,0.663,,3,14,1
+accuracy,Temperature-Rain,Chronos-T5 (Base),WQL,0.669,,3,14,2
+accuracy,Temperature-Rain,Chronos-T5 (Small),WQL,0.685,,3,14,3
+accuracy,Temperature-Rain,Chronos-T5 (Mini),WQL,0.704,,3,14,4
+accuracy,Uber TLC (Daily),Chronos-T5 (Large),WQL,0.096,,3,15,1
+accuracy,Uber TLC (Daily),Chronos-T5 (Base),WQL,0.097,,3,15,2
+accuracy,Uber TLC (Daily),Chronos-T5 (Small),WQL,0.100,,3,15,3
+accuracy,Uber TLC (Daily),Chronos-T5 (Mini),WQL,0.105,,3,15,4
+accuracy,Uber TLC (Hourly),Chronos-T5 (Large),WQL,0.153,,3,16,1
+accuracy,Uber TLC (Hourly),Chronos-T5 (Base),WQL,0.153,,3,16,2
+accuracy,Uber TLC (Hourly),Chronos-T5 (Small),WQL,0.155,,3,16,3
+accuracy,Uber TLC (Hourly),Chronos-T5 (Mini),WQL,0.161,,3,16,4
+accuracy,Agg. Relative Score,Chronos-T5 (Large),WQL,0.564,,3,17,1
+accuracy,Agg. Relative Score,Chronos-T5 (Base),WQL,0.580,,3,17,2
+accuracy,Agg. Relative Score,Chronos-T5 (Small),WQL,0.603,,3,17,3
+accuracy,Agg. Relative Score,Chronos-T5 (Mini),WQL,0.598,,3,17,4
+accuracy,Avg. Rank,Chronos-T5 (Large),WQL,3.400,,3,18,1
+accuracy,Avg. Rank,Chronos-T5 (Base),WQL,4.667,,3,18,2
+accuracy,Avg. Rank,Chronos-T5 (Small),WQL,6.200,,3,18,3
+accuracy,Avg. Rank,Chronos-T5 (Mini),WQL,6.067,,3,18,4
diff --git a/result/per_paper/2403.07815/components_architecture.csv b/result/per_paper/2403.07815/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..829977bced16679566dfc569f76e7244358221d3
--- /dev/null
+++ b/result/per_paper/2403.07815/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Tokenizer,A component that scales and quantizes time series values into a fixed vocabulary of tokens for input to the language model.,proposed_here,,The input time series is scaled and quantized to obtain a sequence of tokens.
+Language Model (Transformer-based),"A transformer-based architecture (e.g., encoder-decoder or decoder-only) trained on tokenized time series using cross-entropy loss.",reused_cited,"Raffel et al., 2020 (T5 family models)",We pretrained CHRONOS models based on the T5 family [...] and trains existing transformer-based language model architectures on these tokenized time series via the cross-entropy loss.
+TSMixup,A data augmentation technique that generates new time series by convex combinations of base time series from different datasets.,proposed_here,,TSMixup randomly samples a set of base time series [...] and generates new time series based on a convex combination of them.
+KernelSynth,A synthetic time series generation method using Gaussian processes to compose random kernel functions.,proposed_here,,KernelSynth uses Gaussian processes to generate synthetic time series by randomly composing kernel functions.
+Autoregressive Sampling,An inference mechanism that samples tokens from the language model and maps them back to numerical values for forecasting.,proposed_here,,"During inference, we autoregressively sample tokens from the model and map them back to numerical values."
diff --git a/result/per_paper/2403.07815/computational.csv b/result/per_paper/2403.07815/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2403.07815/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2403.20150/accuracy_efficiency.csv b/result/per_paper/2403.20150/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2403.20150/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2403.20150/accuracy_efficiency_traced.csv b/result/per_paper/2403.20150/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2403.20150/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2403.20150/components_architecture.csv b/result/per_paper/2403.20150/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..8b22750924564ca3b1a913462c37bc5c741632cb
--- /dev/null
+++ b/result/per_paper/2403.20150/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+TFB Benchmark Framework,"An automated benchmarking system for Time Series Forecasting (TSF) methods, addressing dataset coverage, method bias, and evaluation pipeline consistency.",proposed_here,"Qiu et al., 2023","We propose TFB, an automated benchmark for Time Series Forecasting (TSF) methods... to address shortcomings related to datasets, comparison methods, and evaluation pipelines."
+Multidomain Dataset Collection,"A curated collection of datasets spanning 10 domains (traffic, electricity, energy, environment, nature, economic, stock markets, banking, health, web) for comprehensive TSF evaluation.",proposed_here,"Qiu et al., 2023","We include datasets from 10 different domains: traffic, electricity, energy, the environment, nature, economic, stock markets, banking, health, and the web."
+Time Series Characterization,A method to ensure dataset comprehensiveness by analyzing characteristics like variability and domain-specific patterns.,proposed_here,"Qiu et al., 2023",We provide a time series characterization to ensure that the selected datasets are comprehensive.
+Diverse Method Inclusion,"A repository of 21 UTSF and 14 MTSF methods spanning statistical, machine learning, and deep learning approaches for unbiased comparison.",proposed_here,"Qiu et al., 2023","We include a diverse range of methods, including statistical learning, machine learning, and deep learning methods."
+Flexible Evaluation Pipeline,"A modular, scalable pipeline enabling fair comparisons through customizable evaluation strategies and metrics.",proposed_here,"Qiu et al., 2023",TFB features a flexible and scalable pipeline that eliminates biases.
+Normalized Performance Metrics,"Standardized metrics (e.g., normalized RMSE) for cross-dataset comparison of forecasting methods.",proposed_here,"Qiu et al., 2023",Box plot of the variations in normalized values of characteristics across the multivariate datasets in the TFB and TSlib.
diff --git a/result/per_paper/2403.20150/computational.csv b/result/per_paper/2403.20150/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..34e87efc59f447906570765e3b9cfedc70437d98
--- /dev/null
+++ b/result/per_paper/2403.20150/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,CPU and GPU,,stated,Compatibility with CPU and GPU hardware enables evaluation in different computing environments.
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,32,,stated,"Initially, the batch size is set to 32, with the option to reduce it by half (to a minimum of 8) in case of an Out-Of-Memory (OOM) situation."
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2405.14755/accuracy_efficiency.csv b/result/per_paper/2405.14755/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2405.14755/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2405.14755/accuracy_efficiency_traced.csv b/result/per_paper/2405.14755/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2405.14755/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2405.14755/components_architecture.csv b/result/per_paper/2405.14755/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d7d9290c57bb0e376277cc1a25d736aae424c328
--- /dev/null
+++ b/result/per_paper/2405.14755/components_architecture.csv
@@ -0,0 +1,4 @@
+component,what_it_is,provenance,citation,evidence
+Time-Series-to-Text Conversion Module,A component that transforms time series data into text format compatible with large language models (LLMs).,proposed_here,"Alnegheimish et al., 2022","Our framework includes a time-series-to-text conversion module, as well as end-to-end pipelines that prompt language models to perform time series anomaly detection."
+PROMPTER Pipeline,A prompting method that directly asks LLMs to identify anomalous elements in the input sequence.,proposed_here,"Alnegheimish et al., 2022","First, we present a prompt-based detection method that directly asks a language model to indicate which elements of the input are anomalies."
+DETECTOR Pipeline,A method leveraging LLMs' forecasting capabilities to detect anomalies by comparing original and forecasted signals.,proposed_here,"Alnegheimish et al., 2022","Second, we leverage the forecasting capability of a large language model to guide the anomaly detection process."
diff --git a/result/per_paper/2405.14755/computational.csv b/result/per_paper/2405.14755/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..23af9ad88900592e11de61c32a2b2e2ff133ac06
--- /dev/null
+++ b/result/per_paper/2405.14755/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,The paper does not explicitly report hardware_type for SigLLM's own model.
+num_devices,,,not_reported,The paper does not explicitly report num_devices for SigLLM's own model.
+training_cost,,,not_reported,The paper does not explicitly report training_cost for SigLLM's own model.
+training_batch_size,,,not_reported,The paper does not explicitly report training_batch_size for SigLLM's own model.
+training_steps_or_epochs,,,not_reported,The paper does not explicitly report training_steps_or_epochs for SigLLM's own model.
+precision,,,not_reported,The paper does not explicitly report precision for SigLLM's own model.
+inference_latency,,,not_reported,The paper does not explicitly report inference_latency for SigLLM's own model.
+inference_throughput,,,not_reported,The paper does not explicitly report inference_throughput for SigLLM's own model.
+peak_memory,,,not_reported,The paper does not explicitly report peak_memory for SigLLM's own model.
+flops_or_macs,,,not_reported,The paper does not explicitly report flops_or_macs for SigLLM's own model.
+num_inference_samples,,,not_reported,The paper does not explicitly report num_inference_samples for SigLLM's own model.
+params,,,not_reported,The paper does not explicitly report params for SigLLM's own model.
+context_lengths_evaluated,,,not_reported,The paper does not explicitly report context_lengths_evaluated for SigLLM's own model.
+horizon_lengths_evaluated,,,not_reported,The paper does not explicitly report horizon_lengths_evaluated for SigLLM's own model.
+inference_batch_size,,,not_reported,The paper does not explicitly report inference_batch_size for SigLLM's own model.
diff --git a/result/per_paper/2405.15273/accuracy_efficiency.csv b/result/per_paper/2405.15273/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b775d30745648d79296646b2a26ee055b32758c8
--- /dev/null
+++ b/result/per_paper/2405.15273/accuracy_efficiency.csv
@@ -0,0 +1,15 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,D3R,DADA,F1,78.02,
+accuracy,GPT4TS,DADA,F1,83.13,
+accuracy,ModernTCN,DADA,F1,83.16,
+accuracy,DADA,DADA,F1,84.57,
+accuracy,D3R,DADA,F1,77.02,
+accuracy,GPT4TS,DADA,F1,77.23,
+accuracy,ModernTCN,DADA,F1,77.17,
+accuracy,DADA,DADA,F1,78.48,
+accuracy,D3R,DADA,F1,80.29,
+accuracy,GPT4TS,DADA,F1,81.44,
+accuracy,ModernTCN,DADA,F1,79.59,
+accuracy,DADA,DADA,F1,82.26,
+accuracy,single 256,DADA-256,F1,256,
+accuracy,MoE 256,DADA-256,F1,256,
diff --git a/result/per_paper/2405.15273/accuracy_efficiency_traced.csv b/result/per_paper/2405.15273/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..45c9b68cd16f6f85252086b5738850d429f0e1cb
--- /dev/null
+++ b/result/per_paper/2405.15273/accuracy_efficiency_traced.csv
@@ -0,0 +1,15 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,D3R,DADA,F1,78.02,,1,2,1
+accuracy,GPT4TS,DADA,F1,83.13,,1,3,1
+accuracy,ModernTCN,DADA,F1,83.16,,1,4,1
+accuracy,DADA,DADA,F1,84.57,,1,5,1
+accuracy,D3R,DADA,F1,77.02,,1,7,1
+accuracy,GPT4TS,DADA,F1,77.23,,1,8,1
+accuracy,ModernTCN,DADA,F1,77.17,,1,9,1
+accuracy,DADA,DADA,F1,78.48,,1,10,1
+accuracy,D3R,DADA,F1,80.29,,1,12,1
+accuracy,GPT4TS,DADA,F1,81.44,,1,13,1
+accuracy,ModernTCN,DADA,F1,79.59,,1,14,1
+accuracy,DADA,DADA,F1,82.26,,1,15,1
+accuracy,single 256,DADA-256,F1,256,,5,2,0
+accuracy,MoE 256,DADA-256,F1,256,,5,3,0
diff --git a/result/per_paper/2405.15273/components_architecture.csv b/result/per_paper/2405.15273/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ad440e73744bb0147bc72cfd1a55b1013c3b9e0d
--- /dev/null
+++ b/result/per_paper/2405.15273/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Adaptive Bottlenecks,"A mechanism enabling flexible selection of information compression levels tailored to different datasets, balancing data fidelity and noise suppression.",proposed_here,"Shentu et al., 2024",enables flexible selection of bottlenecks based on different data and explicitly enhances clear differentiation between normal and abnormal series.
+Dual Adversarial Decoders,Two adversarial decoders designed to enhance the model's ability to distinguish between normal and abnormal time series patterns.,proposed_here,"Shentu et al., 2024","Dual Adversarial Decoders (DADA), which enables flexible selection of bottlenecks based on different data and explicitly enhances clear differentiation between normal and abnormal series."
+Autoencoder,"A foundational component for learning normal data patterns through reconstruction, leveraging the scarcity of anomalies.",reused_cited,"Chalapathy & Chawla, 2019","ing the characteristics of normal data with autoencoder, due to the inherently scarce nature of anomalies."
+Information Bottleneck,"A theoretical framework for balancing compact data compression and high-fidelity reconstruction, guiding bottleneck design.",reused_cited,"Kawaguchi et al., 2023",Information bottleneck is considered as a trade-off between compactly compressing the intrinsic information of the original data and high-fidelity reconstruction.
diff --git a/result/per_paper/2405.15273/computational.csv b/result/per_paper/2405.15273/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a1b83e35ce9886f3f669046e30aefb60cd7a846f
--- /dev/null
+++ b/result/per_paper/2405.15273/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware specifications (e.g., GPU/TPU types) are mentioned in the text or tables."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not reported."
+training_cost,,,not_reported,"Training cost (e.g., monetary cost, computational resources) is not mentioned."
+training_batch_size,,,not_reported,Batch size during training is not specified in the text or tables.
+training_steps_or_epochs,,,not_reported,Number of training steps or epochs is not reported.
+precision,,,not_reported,"Precision metric (e.g., F1, P, R) is not explicitly reported for DADA in the text; only F1 scores are provided in tables."
+inference_latency,0.081,seconds,stated,Inference Time(s) for DADA is explicitly listed as 0.081 in the 'Methods' table.
+inference_throughput,,,not_reported,"Inference throughput (e.g., samples/second) is not mentioned."
+peak_memory,,,not_reported,Peak memory usage during training or inference is not reported.
+flops_or_macs,,,not_reported,FLOPs or MACs (computational complexity) are not mentioned.
+num_inference_samples,,,not_reported,Number of inference samples tested is not specified.
+params,,,not_reported,Number of model parameters is not reported.
+context_lengths_evaluated,,,not_reported,"Context lengths (e.g., sequence lengths) evaluated during testing are not mentioned."
+horizon_lengths_evaluated,,,not_reported,"Horizon lengths (e.g., prediction horizons) evaluated are not specified."
+inference_batch_size,,,not_reported,Inference batch size is not reported in the text or tables.
diff --git a/result/per_paper/2405.15370/accuracy_efficiency.csv b/result/per_paper/2405.15370/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5560c002a9d274228e4cbbc268e379c536173111
--- /dev/null
+++ b/result/per_paper/2405.15370/accuracy_efficiency.csv
@@ -0,0 +1,4 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Llama-3-70B,LLM accurate interpretable TSAD,F1,-3,
+accuracy,GPT-3.5,LLM accurate interpretable TSAD,F1,-3.5,
+accuracy,GPT-4,LLM accurate interpretable TSAD,F1,-4,
diff --git a/result/per_paper/2405.15370/accuracy_efficiency_traced.csv b/result/per_paper/2405.15370/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..0727f90bd4abebb8de5df14c25181107fd6e3a8e
--- /dev/null
+++ b/result/per_paper/2405.15370/accuracy_efficiency_traced.csv
@@ -0,0 +1,4 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Llama-3-70B,LLM accurate interpretable TSAD,F1,-3,,4,2,0
+accuracy,GPT-3.5,LLM accurate interpretable TSAD,F1,-3.5,,4,3,0
+accuracy,GPT-4,LLM accurate interpretable TSAD,F1,-4,,4,4,0
diff --git a/result/per_paper/2405.15370/components_architecture.csv b/result/per_paper/2405.15370/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..8ce41b31c6205c74c4b575061b39eee6c60158b6
--- /dev/null
+++ b/result/per_paper/2405.15370/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+LLM (Large Language Model),Pretrained language model used as the core for anomaly detection and explanation generation.,reused_cited,1,"LLMs are pretrained with diverse textual data, including numerical time series [22], and excel in in-context learning (ICL) with only a few-shot examples [7, 23, 32]."
+Time Series Segment Retrieval Module,Component that retrieves and leverages positive and negative similar time series segments for few-shot anomaly detection.,proposed_here,,LLMAD innovatively applies LLMs for few-shot anomaly detection by retrieving and leveraging both positive and negative similar time series segments.
+Anomaly Detection Chain-of-Thought (AnoCoT) Prompting,Prompting strategy that mimics expert logic for decision-making and provides interpretable explanations.,proposed_here,,LLMAD employs the Anomaly Detection Chain-of-Thought (AnoCoT) approach to mimic expert logic for its decision-making process.
+Explanation Generation Module,Component that integrates domain-specific knowledge into prompts to generate human-readable explanations for anomaly detections.,proposed_here,,"LLMs can provide anomaly points, anomaly types, alarm levels, and explanations by integrating relevant knowledge directly into the prompts, such as data background information and anomaly detection rules."
+In-Context Learning (ICL) Integration,Mechanism to activate in-context learning by inputting retrieved historical patterns (normal/abnormal) for few-shot prediction.,proposed_here,,"LLMAD retrieves both normal and abnormal patterns from history as input to activate the In-Context Learning (ICL) [5], thereby enabling accurate anomaly prediction with few shots."
diff --git a/result/per_paper/2405.15370/computational.csv b/result/per_paper/2405.15370/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..8440faaddb88bba538ff162066d6719e156ad0b0
--- /dev/null
+++ b/result/per_paper/2405.15370/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,No hardware specifications mentioned for the model.
+num_devices,,,not_reported,No information on number of devices used for training or inference.
+training_cost,,,not_reported,"No mention of computational cost (e.g., monetary, energy, or resource expenditure)."
+training_batch_size,,,not_reported,Batch size details for training are not provided.
+training_steps_or_epochs,,,not_reported,No information on training duration (steps/epochs).
+precision,,,not_reported,"Precision settings (e.g., FP16, FP32) are not mentioned."
+inference_latency,,,not_reported,Latency metrics for inference are not reported.
+inference_throughput,,,not_reported,Throughput during inference is not specified.
+peak_memory,,,not_reported,Peak memory usage during training or inference is not mentioned.
+flops_or_macs,,,not_reported,No FLOPs or MACs calculations provided.
+num_inference_samples,,,not_reported,Number of samples used for inference evaluation is not explicitly stated.
+params,,,not_reported,Model parameter count is not reported.
+context_lengths_evaluated,,,not_reported,Context length details for evaluation are not provided.
+horizon_lengths_evaluated,,,not_reported,Horizon length details for evaluation are not mentioned.
+inference_batch_size,,,not_reported,Inference batch size is not specified.
diff --git a/result/per_paper/2407.07311/accuracy_efficiency.csv b/result/per_paper/2407.07311/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b138ff116003f87eaa3b1a18b11fd5c0c1aab10d
--- /dev/null
+++ b/result/per_paper/2407.07311/accuracy_efficiency.csv
@@ -0,0 +1,22 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,ViTime-0.5,MSE,0.428,0.009
+accuracy,ETTh1,ViTime-0.7,MSE,0.456,0.005
+accuracy,Mean,ViTime-0.5,MSE,0.515,
+accuracy,ETTh2,ViTime-0.5,MSE,0.332,0.003
+accuracy,ETTh2,ViTime-0.7,MSE,0.342,0.003
+accuracy,Mean,ViTime-0.5,MSE,0.345,
+accuracy,ETTm1,ViTime-0.5,MSE,0.448,0.007
+accuracy,ETTm1,ViTime-0.7,MSE,0.418,0.003
+accuracy,Mean,ViTime-0.5,MSE,0.477,
+accuracy,ETTm2,ViTime-0.5,MSE,0.285,0.006
+accuracy,ETTm2,ViTime-0.7,MSE,0.292,0.007
+accuracy,Mean,ViTime-0.5,MSE,0.293,
+accuracy,Electricity,ViTime-0.5,MSE,0.377,0.004
+accuracy,Electricity,ViTime-0.7,MSE,0.357,0.006
+accuracy,Mean,ViTime-0.5,MSE,0.287,
+accuracy,Traffic,ViTime-0.5,MSE,0.495,0.011
+accuracy,Traffic,ViTime-0.7,MSE,0.503,0.026
+accuracy,Mean,ViTime-0.5,MSE,0.849,
+accuracy,Weather,ViTime-0.5,MSE,0.230,0.005
+accuracy,Weather,ViTime-0.7,MSE,0.240,0.010
+accuracy,Mean,ViTime-0.5,MSE,0.272,
diff --git a/result/per_paper/2407.07311/accuracy_efficiency_traced.csv b/result/per_paper/2407.07311/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4c1fa127c72c6247fa2798ce3b256d925158156c
--- /dev/null
+++ b/result/per_paper/2407.07311/accuracy_efficiency_traced.csv
@@ -0,0 +1,22 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,ViTime-0.5,MSE,0.428,0.009,1,2,7
+accuracy,ETTh1,ViTime-0.7,MSE,0.456,0.005,1,2,9
+accuracy,Mean,ViTime-0.5,MSE,0.515,,1,6,7
+accuracy,ETTh2,ViTime-0.5,MSE,0.332,0.003,1,7,7
+accuracy,ETTh2,ViTime-0.7,MSE,0.342,0.003,1,7,9
+accuracy,Mean,ViTime-0.5,MSE,0.345,,1,11,7
+accuracy,ETTm1,ViTime-0.5,MSE,0.448,0.007,1,12,7
+accuracy,ETTm1,ViTime-0.7,MSE,0.418,0.003,1,12,9
+accuracy,Mean,ViTime-0.5,MSE,0.477,,1,16,7
+accuracy,ETTm2,ViTime-0.5,MSE,0.285,0.006,1,17,7
+accuracy,ETTm2,ViTime-0.7,MSE,0.292,0.007,1,17,9
+accuracy,Mean,ViTime-0.5,MSE,0.293,,1,21,7
+accuracy,Electricity,ViTime-0.5,MSE,0.377,0.004,1,22,7
+accuracy,Electricity,ViTime-0.7,MSE,0.357,0.006,1,22,9
+accuracy,Mean,ViTime-0.5,MSE,0.287,,1,26,7
+accuracy,Traffic,ViTime-0.5,MSE,0.495,0.011,1,27,7
+accuracy,Traffic,ViTime-0.7,MSE,0.503,0.026,1,27,9
+accuracy,Mean,ViTime-0.5,MSE,0.849,,1,31,7
+accuracy,Weather,ViTime-0.5,MSE,0.230,0.005,1,32,7
+accuracy,Weather,ViTime-0.7,MSE,0.240,0.010,1,32,9
+accuracy,Mean,ViTime-0.5,MSE,0.272,,1,36,7
diff --git a/result/per_paper/2407.07311/components_architecture.csv b/result/per_paper/2407.07311/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..95d6271c352a669449bfd9bb9f46e4558d6314fd
--- /dev/null
+++ b/result/per_paper/2407.07311/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Binary Image-Based Time Series Metric Space,"A formally defined metric space where numerical time series are transformed into binary images, enabling operations based on pixel spatial patterns for forecasting.",proposed_here,,ViTime fundamentally shifts TSF from numerical fitting to operations based on a binary image-based time series metric space.
+RealTS (Real-Time Series) Data Generation Algorithm,An innovative synthesis algorithm that generates diverse and realistic training samples by categorizing time series knowledge into 'trend' and 'periodicity' during ViTime training.,proposed_here,,RealTS categorizes foundational knowledge of time series analysis into 'trend' and 'periodicity' and synthesizes training data during the training of ViTime.
+Quantization-Induced System Error Bounds,"Theoretical analysis of errors introduced during the transformation of numerical time series into binary images, providing bounds for system accuracy.",proposed_here,,We provide detailed theoretical analyses of quantization-induced errors and establish principled guidelines for optimal parameter settings.
+Principled Strategies for Optimal Parameter Selection,Guidelines for balancing computational complexity and prediction accuracy through systematic parameter tuning in the binary image metric space.,proposed_here,,We provide detailed theoretical analyses of quantization-induced errors and establish principled guidelines for optimal parameter settings.
+Visual Space Data Operation Paradigm,A framework for solving time series forecasting tasks (point and probabilistic) by operating directly on binary image representations of time series data.,proposed_here,,ViTime operates by transforming numerical time series into binary images... solving TSF tasks in binary image space.
diff --git a/result/per_paper/2407.07311/computational.csv b/result/per_paper/2407.07311/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..34842385e0b73058505b1a7e20bfc55fd8186fd3
--- /dev/null
+++ b/result/per_paper/2407.07311/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware type (e.g., GPU/TPU) is explicitly mentioned for ViTime."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training/inference is not stated."
+training_cost,,,not_reported,"Training cost (e.g., monetary or computational) is not mentioned."
+training_batch_size,,,not_reported,The training batch size is not explicitly reported.
+training_steps_or_epochs,,,not_reported,The number of training steps or epochs is not stated.
+precision,,,not_reported,"Precision (e.g., 32-bit vs. 16-bit) is not mentioned."
+inference_latency,,,not_reported,"Inference latency (e.g., time per sample) is not reported."
+inference_throughput,,,not_reported,"Inference throughput (e.g., samples/second) is not mentioned."
+peak_memory,,,not_reported,"Peak memory usage (e.g., GPU memory) is not stated."
+flops_or_macs,,,not_reported,FLOPs or MACs (computational complexity) are not reported.
+num_inference_samples,,,not_reported,The number of inference samples tested is not explicitly mentioned.
+params,,,not_reported,The number of model parameters is not stated.
+context_lengths_evaluated,512,timesteps,stated,The maximum lookback window T = 512 is explicitly mentioned as a default parameter.
+horizon_lengths_evaluated,"96, 192, 336, 720",timesteps,stated,"The models are evaluated on future sequence lengths of 96, 192, 336, and 720 timesteps."
+inference_batch_size,,,not_reported,The inference batch size is not explicitly reported.
diff --git a/result/per_paper/2407.07874/accuracy_efficiency.csv b/result/per_paper/2407.07874/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d5ea50bf664b408420c70e62d1d902497d33d17f
--- /dev/null
+++ b/result/per_paper/2407.07874/accuracy_efficiency.csv
@@ -0,0 +1,17 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,Toto,MAE,0.389,
+accuracy,MSE,Toto,MAE,0.400,
+accuracy,ETTh2,Toto,MAE,0.261,
+accuracy,MSE,Toto,MAE,0.341,
+accuracy,ETTm1,Toto,MAE,0.375,
+accuracy,MSE,Toto,MAE,0.448,
+accuracy,ETTm2,Toto,MAE,0.319,
+accuracy,MSE,Toto,MAE,0.300,
+accuracy,Electricity,Toto,MAE,0.246,
+accuracy,MSE,Toto,MAE,0.233,
+accuracy,Weather,Toto,MAE,0.284,
+accuracy,MSE,Toto,MAE,0.242,
+accuracy,Mean,Toto,MAE,0.312,
+accuracy,MSE,Toto,MAE,0.328,
+accuracy,sMAPE,Toto,sMAPE,0.672,
+accuracy,sMdAPE,Toto,sMAPE,0.318,
diff --git a/result/per_paper/2407.07874/accuracy_efficiency_traced.csv b/result/per_paper/2407.07874/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..cd66c172bd7fc04899330e14343566b358e728a2
--- /dev/null
+++ b/result/per_paper/2407.07874/accuracy_efficiency_traced.csv
@@ -0,0 +1,17 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,Toto,MAE,0.389,,1,2,2
+accuracy,MSE,Toto,MAE,0.400,,1,3,2
+accuracy,ETTh2,Toto,MAE,0.261,,1,4,2
+accuracy,MSE,Toto,MAE,0.341,,1,5,2
+accuracy,ETTm1,Toto,MAE,0.375,,1,6,2
+accuracy,MSE,Toto,MAE,0.448,,1,7,2
+accuracy,ETTm2,Toto,MAE,0.319,,1,8,2
+accuracy,MSE,Toto,MAE,0.300,,1,9,2
+accuracy,Electricity,Toto,MAE,0.246,,1,10,2
+accuracy,MSE,Toto,MAE,0.233,,1,11,2
+accuracy,Weather,Toto,MAE,0.284,,1,12,2
+accuracy,MSE,Toto,MAE,0.242,,1,13,2
+accuracy,Mean,Toto,MAE,0.312,,1,14,2
+accuracy,MSE,Toto,MAE,0.328,,1,15,2
+accuracy,sMAPE,Toto,sMAPE,0.672,,2,1,1
+accuracy,sMdAPE,Toto,sMAPE,0.318,,2,2,1
diff --git a/result/per_paper/2407.07874/components_architecture.csv b/result/per_paper/2407.07874/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..67985ac34b7eadb0c65b58aae9a8c2e0ad5ba3ee
--- /dev/null
+++ b/result/per_paper/2407.07874/components_architecture.csv
@@ -0,0 +1,11 @@
+component,what_it_is,provenance,citation,evidence
+Patch Embedding,A layer that projects multivariate time series data into fixed-size patches for input to the transformer stack.,reused_cited,PatchTST [14],non-overlapping patch projections... popularized in the time series context by PatchTST [14].
+Space-Wise Multi-Head Attention (MHA),"An attention mechanism that groups multivariate time series features efficiently, reducing computational overhead.",proposed_here,,Proportional factorized space-time attention: We introduce an advanced attention mechanism that allows for efficient grouping of multivariate time series features...
+RMSNorm,A normalization layer used in transformer blocks for improved training stability and efficiency.,reused_cited,RMSNorm [28],"We use techniques from some of the latest large language model (LLM) architectures, including... RMSNorm [28]."
+Space-Wise Block,"A transformer block containing space-wise MHA followed by feed-forward layers, designed for cross-series feature interactions.",proposed_here,,Each segment of the transformer consists of one space-wise transformer block followed by N time-wise blocks.
+Feed Forward (SwiGLU),A feed-forward network using the SwiGLU activation function for improved model expressiveness.,reused_cited,SwiGLU [29],"We use techniques from some of the latest large language model (LLM) architectures, including... SwiGLU feed-forward layers [29]."
+Time-Wise Multi-Head Attention (MHA) + RoPE,An attention mechanism with rotary position embeddings (RoPE) for temporal dependencies and causal prediction.,proposed_here,,Time-Wise MHA + RoPE... The causal next-patch prediction task also simplifies the pre-training process.
+Time-Wise Block,"A transformer block containing time-wise MHA with RoPE and feed-forward layers, designed for temporal pattern capture.",proposed_here,,Each segment of the transformer consists of one space-wise transformer block followed by N time-wise blocks.
+Student-T Mixture Model Head,A probabilistic output layer using a mixture of Student-T distributions for robust uncertainty estimation.,proposed_here,,Student-T mixture model head: This novel use of a probabilistic model... provides superior performance over traditional approaches.
+Mixture Weights,Parameters that determine the contribution of each Student-T component in the mixture distribution.,proposed_here,,Flattened transformer outputs are projected to form the parameters of the Student-T mixture model (SMM) head.
+Mixture Distribution,The final probabilistic output combining multiple Student-T distributions for forecasting.,proposed_here,,"The final outputs are the forecasts for the input series, shifted P steps (the patch width) into the future."
diff --git a/result/per_paper/2407.07874/computational.csv b/result/per_paper/2407.07874/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..8dd9cb60aee4cabf66d6aaa6a7890c24fe307bd3
--- /dev/null
+++ b/result/per_paper/2407.07874/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No specific hardware (e.g., GPU/TPU) is mentioned."
+num_devices,,,not_reported,No information on the number of devices used for training or inference.
+training_cost,,,not_reported,"No explicit mention of computational cost (e.g., monetary or resource cost)."
+training_batch_size,,,not_reported,Batch size is not specified in the text.
+training_steps_or_epochs,,,not_reported,No specific number of training steps or epochs is reported.
+precision,,,not_reported,"No mention of numerical precision (e.g., 32-bit, 16-bit)."
+inference_latency,,,not_reported,Latency metrics are not provided.
+inference_throughput,,,not_reported,Throughput metrics are not provided.
+peak_memory,,,not_reported,No memory usage (peak or otherwise) is reported.
+flops_or_macs,,,not_reported,No FLOPs or MACs are mentioned.
+num_inference_samples,100,samples,stated,"The text states: 'we vary the number of samples generated based on the cardinality and length of the dataset, with a minimum of 100 samples.'"
+params,,,not_reported,Model parameter count is not specified.
+context_lengths_evaluated,512,steps,stated,The text states: 'we use a historical context window of 512 steps.'
+horizon_lengths_evaluated,365,steps,stated,"The text states: 'we evaluate using a prediction length of 365, the maximum forecast window available.'"
+inference_batch_size,,,not_reported,Inference batch size is not mentioned.
diff --git a/result/per_paper/2408.17253/accuracy_efficiency.csv b/result/per_paper/2408.17253/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..830bb1eda00c1d8618149ce1c65a04eec278f4e7
--- /dev/null
+++ b/result/per_paper/2408.17253/accuracy_efficiency.csv
@@ -0,0 +1,90 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,VisionTS,MAE,0.456,
+accuracy,avg,VisionTS,MAE,0.691,
+accuracy,ETTh2,VisionTS,MAE,0.374,
+accuracy,avg,VisionTS,MAE,0.605,
+accuracy,ETTm1,VisionTS,MAE,0.404,
+accuracy,avg,VisionTS,MAE,0.411,
+accuracy,ETTm2,VisionTS,MAE,0.269,
+accuracy,avg,VisionTS,MAE,0.316,
+accuracy,Electricity,VisionTS,MAE,0.237,
+accuracy,avg,VisionTS,MAE,0.180,
+accuracy,Weather,VisionTS,MAE,0.215,
+accuracy,avg,VisionTS,MAE,0.241,
+accuracy,M1 Monthly,VisionTS,MAE,1987.69,
+accuracy,M3 Monthly,VisionTS,MAE,737.93,
+accuracy,M3 Other,VisionTS,MAE,315.85,
+accuracy,M4 Monthly,VisionTS,MAE,666.54,
+accuracy,M4 Weekly,VisionTS,MAE,404.23,
+accuracy,M4 Daily,VisionTS,MAE,215.63,
+accuracy,M4 Hourly,VisionTS,MAE,288.37,
+accuracy,Tourism Quarterly,VisionTS,MAE,12931.88,
+accuracy,Tourism Monthly,VisionTS,MAE,2560.19,
+accuracy,CIF 2016,VisionTS,MAE,570907.24,
+accuracy,Aus. Elec. Demand,VisionTS,MAE,237.44,
+accuracy,Bitcoin,VisionTS,MAE,2.33,
+accuracy,Pedestrian Counts,VisionTS,MAE,52.01,
+accuracy,Vehicle Trips,VisionTS,MAE,22.08,
+accuracy,KDD cup,VisionTS,MAE,38.16,
+accuracy,Weather,VisionTS,MAE,2.06,
+accuracy,NN5 Daily,VisionTS,MAE,3.51,
+accuracy,NN5 Weekly,VisionTS,MAE,14.67,
+accuracy,Carparts,VisionTS,MAE,0.58,
+accuracy,FRED-MD,VisionTS,MAE,1893.67,
+accuracy,Traffic Hourly,VisionTS,MAE,0.01,
+accuracy,Traffic Weekly,VisionTS,MAE,1.14,
+accuracy,Rideshare,VisionTS,MAE,5.92,
+accuracy,Hospital,VisionTS,MAE,19.36,
+accuracy,COVID Deaths,VisionTS,MAE,137.51,
+accuracy,Temperature Rain,VisionTS,MAE,6.37,
+accuracy,Sunspot,VisionTS,MAE,2.81,
+accuracy,Saugeen River Flow,VisionTS,MAE,30.22,
+accuracy,US Births,VisionTS,MAE,519.94,
+accuracy,Normalized MAE,VisionTS,MAE,0.729,
+accuracy,Rank,VisionTS,MAE,2,
+accuracy,ETTh1,VisionTS (Base),MAE,0.375,
+accuracy,ETTh1,VisionTS (Large),MAE,0.402,
+accuracy,ETTh1,VisionTS (Huge),MAE,0.384,
+accuracy,avg,VisionTS (Base),MAE,0.424,
+accuracy,avg,VisionTS (Large),MAE,0.434,
+accuracy,avg,VisionTS (Huge),MAE,0.439,
+accuracy,ETTh2,VisionTS (Base),MAE,0.281,
+accuracy,ETTh2,VisionTS (Large),MAE,0.334,
+accuracy,ETTh2,VisionTS (Huge),MAE,0.277,
+accuracy,avg,VisionTS (Base),MAE,0.379,
+accuracy,avg,VisionTS (Large),MAE,0.346,
+accuracy,avg,VisionTS (Huge),MAE,0.382,
+accuracy,ETTm1,VisionTS (Base),MAE,0.404,
+accuracy,ETTm1,VisionTS (Large),MAE,0.383,
+accuracy,ETTm1,VisionTS (Huge),MAE,0.335,
+accuracy,avg,VisionTS (Base),MAE,0.410,
+accuracy,avg,VisionTS (Large),MAE,0.382,
+accuracy,avg,VisionTS (Huge),MAE,0.388,
+accuracy,ETTm2,VisionTS (Base),MAE,0.205,
+accuracy,ETTm2,VisionTS (Large),MAE,0.282,
+accuracy,ETTm2,VisionTS (Huge),MAE,0.195,
+accuracy,avg,VisionTS (Base),MAE,0.341,
+accuracy,avg,VisionTS (Large),MAE,0.272,
+accuracy,avg,VisionTS (Huge),MAE,0.321,
+accuracy,Electricity,VisionTS (Base),MAE,0.205,
+accuracy,Electricity,VisionTS (Large),MAE,0.299,
+accuracy,Electricity,VisionTS (Huge),MAE,0.158,
+accuracy,avg,VisionTS (Base),MAE,0.320,
+accuracy,avg,VisionTS (Large),MAE,0.188,
+accuracy,avg,VisionTS (Huge),MAE,0.274,
+accuracy,Weather,VisionTS (Base),MAE,0.173,
+accuracy,Weather,VisionTS (Large),MAE,0.212,
+accuracy,Weather,VisionTS (Huge),MAE,0.167,
+accuracy,avg,VisionTS (Base),MAE,0.267,
+accuracy,avg,VisionTS (Large),MAE,0.238,
+accuracy,avg,VisionTS (Huge),MAE,0.261,
+accuracy,Average 1st count,VisionTS (Base),MAE,0.357,
+accuracy,Average 1st count,VisionTS (Large),MAE,0.310,
+accuracy,Average 1st count,VisionTS (Huge),MAE,0.344,
+accuracy,ETTh1,VisionTS,MSE,0.402,
+accuracy,ETTh2,VisionTS,MSE,0.346,
+accuracy,ETTm1,VisionTS,MSE,0.341,
+accuracy,ETTm2,VisionTS,MSE,0.248,
+accuracy,Weather,VisionTS,MSE,0.199,
+accuracy,Traffic,VisionTS,MSE,0.267,
+accuracy,Electricity,VisionTS,MSE,0.233,
diff --git a/result/per_paper/2408.17253/accuracy_efficiency_traced.csv b/result/per_paper/2408.17253/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5e4e9d2ae6c3e7357417179b4fd8f9c5975daa3f
--- /dev/null
+++ b/result/per_paper/2408.17253/accuracy_efficiency_traced.csv
@@ -0,0 +1,90 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,VisionTS,MAE,0.456,,1,4,13
+accuracy,avg,VisionTS,MAE,0.691,,1,8,13
+accuracy,ETTh2,VisionTS,MAE,0.374,,1,9,13
+accuracy,avg,VisionTS,MAE,0.605,,1,13,13
+accuracy,ETTm1,VisionTS,MAE,0.404,,1,14,13
+accuracy,avg,VisionTS,MAE,0.411,,1,18,13
+accuracy,ETTm2,VisionTS,MAE,0.269,,1,19,13
+accuracy,avg,VisionTS,MAE,0.316,,1,23,13
+accuracy,Electricity,VisionTS,MAE,0.237,,1,24,13
+accuracy,avg,VisionTS,MAE,0.180,,1,28,13
+accuracy,Weather,VisionTS,MAE,0.215,,1,29,13
+accuracy,avg,VisionTS,MAE,0.241,,1,33,13
+accuracy,M1 Monthly,VisionTS,MAE,1987.69,,2,1,1
+accuracy,M3 Monthly,VisionTS,MAE,737.93,,2,2,1
+accuracy,M3 Other,VisionTS,MAE,315.85,,2,3,1
+accuracy,M4 Monthly,VisionTS,MAE,666.54,,2,4,1
+accuracy,M4 Weekly,VisionTS,MAE,404.23,,2,5,1
+accuracy,M4 Daily,VisionTS,MAE,215.63,,2,6,1
+accuracy,M4 Hourly,VisionTS,MAE,288.37,,2,7,1
+accuracy,Tourism Quarterly,VisionTS,MAE,12931.88,,2,8,1
+accuracy,Tourism Monthly,VisionTS,MAE,2560.19,,2,9,1
+accuracy,CIF 2016,VisionTS,MAE,570907.24,,2,10,1
+accuracy,Aus. Elec. Demand,VisionTS,MAE,237.44,,2,11,1
+accuracy,Bitcoin,VisionTS,MAE,2.33,,2,12,1
+accuracy,Pedestrian Counts,VisionTS,MAE,52.01,,2,13,1
+accuracy,Vehicle Trips,VisionTS,MAE,22.08,,2,14,1
+accuracy,KDD cup,VisionTS,MAE,38.16,,2,15,1
+accuracy,Weather,VisionTS,MAE,2.06,,2,16,1
+accuracy,NN5 Daily,VisionTS,MAE,3.51,,2,17,1
+accuracy,NN5 Weekly,VisionTS,MAE,14.67,,2,18,1
+accuracy,Carparts,VisionTS,MAE,0.58,,2,19,1
+accuracy,FRED-MD,VisionTS,MAE,1893.67,,2,20,1
+accuracy,Traffic Hourly,VisionTS,MAE,0.01,,2,21,1
+accuracy,Traffic Weekly,VisionTS,MAE,1.14,,2,22,1
+accuracy,Rideshare,VisionTS,MAE,5.92,,2,23,1
+accuracy,Hospital,VisionTS,MAE,19.36,,2,24,1
+accuracy,COVID Deaths,VisionTS,MAE,137.51,,2,25,1
+accuracy,Temperature Rain,VisionTS,MAE,6.37,,2,26,1
+accuracy,Sunspot,VisionTS,MAE,2.81,,2,27,1
+accuracy,Saugeen River Flow,VisionTS,MAE,30.22,,2,28,1
+accuracy,US Births,VisionTS,MAE,519.94,,2,29,1
+accuracy,Normalized MAE,VisionTS,MAE,0.729,,2,30,1
+accuracy,Rank,VisionTS,MAE,2,,2,31,1
+accuracy,ETTh1,VisionTS (Base),MAE,0.375,,3,2,8
+accuracy,ETTh1,VisionTS (Large),MAE,0.402,,3,2,9
+accuracy,ETTh1,VisionTS (Huge),MAE,0.384,,3,2,10
+accuracy,avg,VisionTS (Base),MAE,0.424,,3,6,8
+accuracy,avg,VisionTS (Large),MAE,0.434,,3,6,9
+accuracy,avg,VisionTS (Huge),MAE,0.439,,3,6,10
+accuracy,ETTh2,VisionTS (Base),MAE,0.281,,3,7,8
+accuracy,ETTh2,VisionTS (Large),MAE,0.334,,3,7,9
+accuracy,ETTh2,VisionTS (Huge),MAE,0.277,,3,7,10
+accuracy,avg,VisionTS (Base),MAE,0.379,,3,11,8
+accuracy,avg,VisionTS (Large),MAE,0.346,,3,11,9
+accuracy,avg,VisionTS (Huge),MAE,0.382,,3,11,10
+accuracy,ETTm1,VisionTS (Base),MAE,0.404,,3,12,8
+accuracy,ETTm1,VisionTS (Large),MAE,0.383,,3,12,9
+accuracy,ETTm1,VisionTS (Huge),MAE,0.335,,3,12,10
+accuracy,avg,VisionTS (Base),MAE,0.410,,3,16,8
+accuracy,avg,VisionTS (Large),MAE,0.382,,3,16,9
+accuracy,avg,VisionTS (Huge),MAE,0.388,,3,16,10
+accuracy,ETTm2,VisionTS (Base),MAE,0.205,,3,17,8
+accuracy,ETTm2,VisionTS (Large),MAE,0.282,,3,17,9
+accuracy,ETTm2,VisionTS (Huge),MAE,0.195,,3,17,10
+accuracy,avg,VisionTS (Base),MAE,0.341,,3,21,8
+accuracy,avg,VisionTS (Large),MAE,0.272,,3,21,9
+accuracy,avg,VisionTS (Huge),MAE,0.321,,3,21,10
+accuracy,Electricity,VisionTS (Base),MAE,0.205,,3,22,8
+accuracy,Electricity,VisionTS (Large),MAE,0.299,,3,22,9
+accuracy,Electricity,VisionTS (Huge),MAE,0.158,,3,22,10
+accuracy,avg,VisionTS (Base),MAE,0.320,,3,26,8
+accuracy,avg,VisionTS (Large),MAE,0.188,,3,26,9
+accuracy,avg,VisionTS (Huge),MAE,0.274,,3,26,10
+accuracy,Weather,VisionTS (Base),MAE,0.173,,3,27,8
+accuracy,Weather,VisionTS (Large),MAE,0.212,,3,27,9
+accuracy,Weather,VisionTS (Huge),MAE,0.167,,3,27,10
+accuracy,avg,VisionTS (Base),MAE,0.267,,3,31,8
+accuracy,avg,VisionTS (Large),MAE,0.238,,3,31,9
+accuracy,avg,VisionTS (Huge),MAE,0.261,,3,31,10
+accuracy,Average 1st count,VisionTS (Base),MAE,0.357,,3,32,8
+accuracy,Average 1st count,VisionTS (Large),MAE,0.310,,3,32,9
+accuracy,Average 1st count,VisionTS (Huge),MAE,0.344,,3,32,10
+accuracy,ETTh1,VisionTS,MSE,0.402,,4,2,5
+accuracy,ETTh2,VisionTS,MSE,0.346,,4,6,5
+accuracy,ETTm1,VisionTS,MSE,0.341,,4,10,5
+accuracy,ETTm2,VisionTS,MSE,0.248,,4,14,5
+accuracy,Weather,VisionTS,MSE,0.199,,4,18,5
+accuracy,Traffic,VisionTS,MSE,0.267,,4,22,5
+accuracy,Electricity,VisionTS,MSE,0.233,,4,26,5
diff --git a/result/per_paper/2408.17253/components_architecture.csv b/result/per_paper/2408.17253/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..8cbabaeddca9db33db34dbb12925e86c086b6755
--- /dev/null
+++ b/result/per_paper/2408.17253/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Visual Masked Autoencoder (MAE),"A pre-trained vision model for image reconstruction, adapted to time series forecasting by reformulating TSF as an image reconstruction task.",reused_cited,"He et al., 2022","We focus on visual masked autoencoder (MAE) $^{1}$, a popular CV foundation model (He et al., 2022) by self-supervised pre-training on ImageNet (Deng et al., 2009)."
+Patch-Level Image Reconstruction,"A method to transform 1D time-series data into 2D matrices via segmentation, then render them as images for reconstruction.",proposed_here,,"We transform 1D time-series data into 2D matrices via segmentation. Then, we render the matrices into images and align the forecasting window with masked image patches."
+Masked Image Patches,"Image patches masked during pre-training of MAE, used to align the forecasting window with time-series data.",reused_cited,"He et al., 2022","As an image reconstruction and completion model, MAE can naturally be a numeric series forecaster."
+Forecasting Window Alignment,A technique to align the time-series forecasting window with masked image patches for zero-shot prediction.,proposed_here,,This allows us to make a zero-shot forecast without further adaptation.
+Image Rendering Module,A component to convert segmented time-series matrices into visual images for MAE processing.,proposed_here,,We render the matrices into images and align the forecasting window with masked image patches.
+Segmentation Layer,A layer to convert 1D time-series data into 2D matrices for image-like processing.,proposed_here,,We transform 1D time-series data into 2D matrices via segmentation.
diff --git a/result/per_paper/2408.17253/computational.csv b/result/per_paper/2408.17253/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..90c8c0f13dd33dec6d7a1023ef8826f536c20178
--- /dev/null
+++ b/result/per_paper/2408.17253/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,not_reported
+num_devices,,,not_reported,not_reported
+training_cost,,,not_reported,not_reported
+training_batch_size,,,not_reported,not_reported
+training_steps_or_epochs,1,epoch,stated,fine-tuning for one epoch
+precision,,,not_reported,not_reported
+inference_latency,,,not_reported,not_reported
+inference_throughput,,,not_reported,not_reported
+peak_memory,,,not_reported,not_reported
+flops_or_macs,,,not_reported,not_reported
+num_inference_samples,,,not_reported,not_reported
+params,,,not_reported,not_reported
+context_lengths_evaluated,"1k, 2k, 3k, 4k",time steps,stated,"Context Length Prediction Length table (1k, 2k, 3k, 4k)"
+horizon_lengths_evaluated,"1k, 2k, 3k, 4k",time steps,stated,"Context Length Prediction Length table (1k, 2k, 3k, 4k)"
+inference_batch_size,,,not_reported,not_reported
diff --git a/result/per_paper/2409.16040/accuracy_efficiency.csv b/result/per_paper/2409.16040/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..189e54342ae9dca6f423093f2e2e5b728c16a2b3
--- /dev/null
+++ b/result/per_paper/2409.16040/accuracy_efficiency.csv
@@ -0,0 +1,50 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,Time-MoEbase,MSE,0.388,
+accuracy,ETTh1,Time-MoElarge,MSE,0.414,
+accuracy,ETTh1,Time-MoEultra,MSE,0.404,
+accuracy,ETTh2,Time-MoEbase,MSE,0.330,
+accuracy,ETTh2,Time-MoElarge,MSE,0.315,
+accuracy,ETTh2,Time-MoEultra,MSE,0.349,
+accuracy,ETTm1,Time-MoEbase,MSE,0.361,
+accuracy,ETTm1,Time-MoElarge,MSE,0.361,
+accuracy,ETTm1,Time-MoEultra,MSE,0.370,
+accuracy,ETTm2,Time-MoEbase,MSE,0.274,
+accuracy,ETTm2,Time-MoElarge,MSE,0.202,
+accuracy,ETTm2,Time-MoEultra,MSE,0.270,
+accuracy,Weather,Time-MoEbase,MSE,0.211,
+accuracy,Global Temp,Time-MoEbase,MSE,0.351,
+accuracy,Global Temp,Time-MoElarge,MSE,0.255,
+accuracy,Global Temp,Time-MoEultra,MSE,0.375,
+accuracy,Average,Time-MoEbase,MSE,0.396,
+accuracy,Average,Time-MoElarge,MSE,0.413,
+accuracy,Average,Time-MoEultra,MSE,0.461,
+accuracy,ETTh1,Time-MoE (Ours),MAE,0.393,
+accuracy,ETTh1,Time-MoE_base,MAE,1,
+accuracy,ETTh1,Time-MoE_large,MAE,96,
+accuracy,ETTh1,Time-MoE_ultra,MAE,0.357,
+accuracy,ETTh2,Time-MoE (Ours),MAE,0.308,
+accuracy,ETTh2,Time-MoE_base,MAE,2,
+accuracy,ETTh2,Time-MoE_large,MAE,96,
+accuracy,ETTh2,Time-MoE_ultra,MAE,0.305,
+accuracy,ETTm1,Time-MoE (Ours),MAE,0.420,
+accuracy,ETTm1,Time-MoE_base,MAE,1,
+accuracy,ETTm1,Time-MoE_large,MAE,96,
+accuracy,ETTm1,Time-MoE_ultra,MAE,0.338,
+accuracy,ETTm2,Time-MoE (Ours),MAE,0.247,
+accuracy,ETTm2,Time-MoE_base,MAE,2,
+accuracy,ETTm2,Time-MoE_large,MAE,96,
+accuracy,ETTm2,Time-MoE_ultra,MAE,0.201,
+accuracy,Weather,Time-MoE (Ours),MAE,0.243,
+accuracy,Weather,Time-MoE_large,MAE,96,
+accuracy,Weather,Time-MoE_ultra,MAE,0.160,
+accuracy,Global Temp,Time-MoE (Ours),MAE,0.308,
+accuracy,Global Temp,Time-MoE_large,MAE,96,
+accuracy,Global Temp,Time-MoE_ultra,MAE,0.211,
+accuracy,Average,Time-MoE (Ours),MAE,0.391,
+accuracy,Average,Time-MoE_large,MAE,0.336,
+accuracy,Average,Time-MoE_ultra,MAE,0.384,
+accuracy,TIME-MoEbase w/ FP32,Time-MoE-base,MSE,32,
+accuracy,"TIME-MOEbase w/ {1,8,32,64}",Time-MoEbase,MSE,1,
+accuracy,"TIME-MOEbase w/ {1,8,32}",Time-MoEbase,MSE,1,
+accuracy,"TIME-MOEbase w/ {1,8}",Time-MoEbase,MSE,1,
+accuracy,TIME-MOEbase w/ {1},Time-MoEbase,MSE,1,
diff --git a/result/per_paper/2409.16040/accuracy_efficiency_traced.csv b/result/per_paper/2409.16040/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2d001feccff2a0bc6e0076e75d5388409b8b460c
--- /dev/null
+++ b/result/per_paper/2409.16040/accuracy_efficiency_traced.csv
@@ -0,0 +1,50 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,Time-MoEbase,MSE,0.388,,1,3,13
+accuracy,ETTh1,Time-MoElarge,MSE,0.414,,1,3,14
+accuracy,ETTh1,Time-MoEultra,MSE,0.404,,1,3,15
+accuracy,ETTh2,Time-MoEbase,MSE,0.330,,1,8,13
+accuracy,ETTh2,Time-MoElarge,MSE,0.315,,1,8,14
+accuracy,ETTh2,Time-MoEultra,MSE,0.349,,1,8,15
+accuracy,ETTm1,Time-MoEbase,MSE,0.361,,1,13,13
+accuracy,ETTm1,Time-MoElarge,MSE,0.361,,1,13,14
+accuracy,ETTm1,Time-MoEultra,MSE,0.370,,1,13,15
+accuracy,ETTm2,Time-MoEbase,MSE,0.274,,1,18,13
+accuracy,ETTm2,Time-MoElarge,MSE,0.202,,1,18,14
+accuracy,ETTm2,Time-MoEultra,MSE,0.270,,1,18,15
+accuracy,Weather,Time-MoEbase,MSE,0.211,,1,23,13
+accuracy,Global Temp,Time-MoEbase,MSE,0.351,,1,28,13
+accuracy,Global Temp,Time-MoElarge,MSE,0.255,,1,28,14
+accuracy,Global Temp,Time-MoEultra,MSE,0.375,,1,28,15
+accuracy,Average,Time-MoEbase,MSE,0.396,,1,33,13
+accuracy,Average,Time-MoElarge,MSE,0.413,,1,33,14
+accuracy,Average,Time-MoEultra,MSE,0.461,,1,33,15
+accuracy,ETTh1,Time-MoE (Ours),MAE,0.393,,2,3,12
+accuracy,ETTh1,Time-MoE_base,MAE,1,,2,3,0
+accuracy,ETTh1,Time-MoE_large,MAE,96,,2,3,1
+accuracy,ETTh1,Time-MoE_ultra,MAE,0.357,,2,3,2
+accuracy,ETTh2,Time-MoE (Ours),MAE,0.308,,2,8,12
+accuracy,ETTh2,Time-MoE_base,MAE,2,,2,8,0
+accuracy,ETTh2,Time-MoE_large,MAE,96,,2,8,1
+accuracy,ETTh2,Time-MoE_ultra,MAE,0.305,,2,8,2
+accuracy,ETTm1,Time-MoE (Ours),MAE,0.420,,2,13,12
+accuracy,ETTm1,Time-MoE_base,MAE,1,,2,13,0
+accuracy,ETTm1,Time-MoE_large,MAE,96,,2,13,1
+accuracy,ETTm1,Time-MoE_ultra,MAE,0.338,,2,13,2
+accuracy,ETTm2,Time-MoE (Ours),MAE,0.247,,2,18,12
+accuracy,ETTm2,Time-MoE_base,MAE,2,,2,18,0
+accuracy,ETTm2,Time-MoE_large,MAE,96,,2,18,1
+accuracy,ETTm2,Time-MoE_ultra,MAE,0.201,,2,18,2
+accuracy,Weather,Time-MoE (Ours),MAE,0.243,,2,23,12
+accuracy,Weather,Time-MoE_large,MAE,96,,2,23,1
+accuracy,Weather,Time-MoE_ultra,MAE,0.160,,2,23,2
+accuracy,Global Temp,Time-MoE (Ours),MAE,0.308,,2,28,12
+accuracy,Global Temp,Time-MoE_large,MAE,96,,2,28,1
+accuracy,Global Temp,Time-MoE_ultra,MAE,0.211,,2,28,2
+accuracy,Average,Time-MoE (Ours),MAE,0.391,,2,33,12
+accuracy,Average,Time-MoE_large,MAE,0.336,,2,33,1
+accuracy,Average,Time-MoE_ultra,MAE,0.384,,2,33,2
+accuracy,TIME-MoEbase w/ FP32,Time-MoE-base,MSE,32,,6,3,0
+accuracy,"TIME-MOEbase w/ {1,8,32,64}",Time-MoEbase,MSE,1,,7,1,0
+accuracy,"TIME-MOEbase w/ {1,8,32}",Time-MoEbase,MSE,1,,7,2,0
+accuracy,"TIME-MOEbase w/ {1,8}",Time-MoEbase,MSE,1,,7,3,0
+accuracy,TIME-MOEbase w/ {1},Time-MoEbase,MSE,1,,7,4,0
diff --git a/result/per_paper/2409.16040/components_architecture.csv b/result/per_paper/2409.16040/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..48774e82946aa9e65a7ad8d6a2d5015b3392a876
--- /dev/null
+++ b/result/per_paper/2409.16040/components_architecture.csv
@@ -0,0 +1,8 @@
+component,what_it_is,provenance,citation,evidence
+Mixture of Experts (MoE) Architecture,A sparse architecture that activates only a subset of networks for each prediction to enhance computational efficiency while maintaining high model capacity.,proposed_here,,"By leveraging a sparse mixture-of-experts (MoE) design, TIME-MOE enhances computational efficiency by activating only a subset of networks for each prediction, reducing computational load while maintaining high model capacity."
+Decoder-Only Transformer Models,A family of transformer models that operate in an autoregressive manner to support flexible forecasting horizons and varying input context lengths.,reused_cited,,TIME-MOE comprises a family of decoder-only transformer models that operate in an autoregressive manner and support flexible forecasting horizons with varying input context lengths.
+Sparse Activation Mechanism,A mechanism that activates only a small subset of parameters during inference to reduce computational overhead while preserving model capacity.,proposed_here,,"processed by a sparse transformer decoder, activating only a small subset of parameters."
+Autoregressive Operation,"A sequential processing mechanism where predictions are generated step-by-step based on previous outputs, enabling support for any forecasting horizon.",reused_cited,,operating in an autoregressive manner to support any forecasting horizon.
+Flexible Forecasting Heads,Multiple output heads designed to enable forecasts across diverse scales and resolutions during inference.,proposed_here,,different forecasting heads are utilized to enable forecasts across diverse scales.
+Point-Wise Tokenization and Encoding,A preprocessing step where input time series data is tokenized and encoded point-wise before being processed by the model.,proposed_here,,input time series is point-wise tokenized and encoded before being processed by a sparse transformer decoder.
+Sparse Transformer Decoder,A decoder component that processes input data with sparse parameter activation to optimize computational efficiency.,proposed_here,,"processed by a sparse transformer decoder, activating only a small subset of parameters."
diff --git a/result/per_paper/2409.16040/computational.csv b/result/per_paper/2409.16040/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..aaa211e33da44877e5588841f4b169f8b3d4e56e
--- /dev/null
+++ b/result/per_paper/2409.16040/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No explicit mention of hardware type (e.g., GPU/TPU) in the text."
+num_devices,,,not_reported,"No mention of the number of devices (e.g., GPUs) used for training or inference."
+training_cost,,,not_reported,"No explicit mention of training cost (e.g., monetary cost, energy consumption)."
+training_batch_size,,,not_reported,No mention of training batch size.
+training_steps_or_epochs,,,not_reported,No mention of training steps or epochs.
+precision,,,not_reported,"No mention of numerical precision (e.g., FP16, FP32)."
+inference_latency,,,not_reported,"No mention of inference latency (e.g., time per sample)."
+inference_throughput,,,not_reported,"No mention of inference throughput (e.g., samples per second)."
+peak_memory,,,not_reported,No mention of peak memory usage during training or inference.
+flops_or_macs,,,not_reported,No mention of FLOPs or MACs (computational operations).
+num_inference_samples,,,not_reported,No mention of the number of inference samples evaluated.
+params,2400000000,parameters,stated,The text states: 'we scale a time series FM up to 2.4 billion parameters.'
+context_lengths_evaluated,,,not_reported,"No explicit mention of evaluated context lengths (e.g., input sequence lengths)."
+horizon_lengths_evaluated,,,not_reported,"No explicit mention of evaluated forecast horizons (e.g., output sequence lengths)."
+inference_batch_size,,,not_reported,No mention of inference batch size.
diff --git a/result/per_paper/2410.04803/accuracy_efficiency.csv b/result/per_paper/2410.04803/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..854aecbd82da0f8d99725eba100ef86759085618
--- /dev/null
+++ b/result/per_paper/2410.04803/accuracy_efficiency.csv
@@ -0,0 +1,6 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ECL,Timer-XL,MSE and MAE,0.265,
+accuracy,ETTh1,Timer-XL,MSE and MAE,0.444,
+accuracy,Traffic,Timer-XL,MSE and MAE,0.298,
+accuracy,Weather,Timer-XL,MSE and MAE,0.291,
+accuracy,Solar-Energy,Timer-XL,MSE and MAE,0.283,
diff --git a/result/per_paper/2410.04803/accuracy_efficiency_traced.csv b/result/per_paper/2410.04803/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..1bdfb4f3bc2385f92f25c76ed98248b91be951cd
--- /dev/null
+++ b/result/per_paper/2410.04803/accuracy_efficiency_traced.csv
@@ -0,0 +1,6 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ECL,Timer-XL,MSE and MAE,0.265,,5,2,10
+accuracy,ETTh1,Timer-XL,MSE and MAE,0.444,,5,3,10
+accuracy,Traffic,Timer-XL,MSE and MAE,0.298,,5,4,10
+accuracy,Weather,Timer-XL,MSE and MAE,0.291,,5,5,10
+accuracy,Solar-Energy,Timer-XL,MSE and MAE,0.283,,5,6,10
diff --git a/result/per_paper/2410.04803/components_architecture.csv b/result/per_paper/2410.04803/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d461031106e1c02a331e692e2b7337d4d0bb85ab
--- /dev/null
+++ b/result/per_paper/2410.04803/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Decoder-only Transformer Architecture,"A causal Transformer model structure that processes varying-length contexts for unified forecasting, enabling autoregressive next token prediction.",reused_cited,"Zhao et al., 2023; Rasul et al., 2023; Ansari et al., 2024","Based on contextual flexibility and autoregressive next token prediction, one model can accommodate varying lookback and prediction lengths (Liu et al., 2024b)."
+TimeAttention,"A novel causal self-attention mechanism tailored for multidimensional time series, capturing intra- and inter-series dependencies with positional awareness and maintaining causality and scalability.",proposed_here,,"We propose TimeAttention to facilitate Transformers on multidimensional time series, presenting Kronecker-based masking mechanism to train time-series Transformers in a channel-dependent approach."
+Deft Position Embedding,"A specialized position embedding for multivariate time series, ensuring temporal causality and variable equivalence through permutation-invariant design.",proposed_here,,"With specialized position embedding for multivariate series, TimeAttention is aware of the chronological order of time points and achieves permutation-equivalence (Zaheer et al., 2017) on variables."
+Flattened Time Series Tokens (Patches),A data representation method where multidimensional time series are flattened into a sequence of tokens (patches) for unified processing.,reused_cited,"Nie et al., 2022; Liu et al., 2023; 2024a","Given the generality across contexts, Timer-XL is a versatile solution for various forecasting tasks."
+Kronecker-based Masking Mechanism,A channel-dependent training mechanism integrated into TimeAttention to enforce causal dependencies in multidimensional time series.,proposed_here,,We propose TimeAttention [...] presenting Kronecker-based masking mechanism to train time-series Transformers in a channel-dependent approach.
diff --git a/result/per_paper/2410.04803/computational.csv b/result/per_paper/2410.04803/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ed2c140e3ffa7b6e14239bfa51f02f5fcdc0ebbf
--- /dev/null
+++ b/result/per_paper/2410.04803/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No explicit mention of hardware (e.g., GPU/TPU) used."
+num_devices,,,not_reported,"No mention of number of devices (e.g., GPUs) used."
+training_cost,,,not_reported,"No explicit cost (e.g., monetary or computational) reported."
+training_batch_size,,,not_reported,Batch size not specified in the text.
+training_steps_or_epochs,,,not_reported,No mention of training steps or epochs.
+precision,,,not_reported,"No mention of precision (e.g., FP16/FP32)."
+inference_latency,,,not_reported,No latency metrics provided.
+inference_throughput,,,not_reported,No throughput metrics provided.
+peak_memory,,,not_reported,No peak memory usage reported.
+flops_or_macs,,,not_reported,"Complexity discussed (e.g., O(N²T²)), but no explicit FLOPs/MACS."
+num_inference_samples,,,not_reported,No number of inference samples specified.
+params,,,not_reported,Model parameter count not mentioned.
+context_lengths_evaluated,3072,tokens,stated,Tables mention 'input-3072-pred-96' for Timer-XL.
+horizon_lengths_evaluated,96,tokens,stated,Tables mention 'input-3072-pred-96' for Timer-XL.
+inference_batch_size,,,not_reported,No inference batch size specified.
diff --git a/result/per_paper/2410.05440/accuracy_efficiency.csv b/result/per_paper/2410.05440/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2410.05440/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2410.05440/accuracy_efficiency_traced.csv b/result/per_paper/2410.05440/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2410.05440/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2410.05440/components_architecture.csv b/result/per_paper/2410.05440/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b44f28f7ab593d252b688e8c35508df1caec5348
--- /dev/null
+++ b/result/per_paper/2410.05440/components_architecture.csv
@@ -0,0 +1 @@
+component,what_it_is,provenance,citation,evidence
diff --git a/result/per_paper/2410.05440/computational.csv b/result/per_paper/2410.05440/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2410.05440/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2410.09385/accuracy_efficiency.csv b/result/per_paper/2410.09385/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6648d015b865121e18f64073b2df2e4838fea188
--- /dev/null
+++ b/result/per_paper/2410.09385/accuracy_efficiency.csv
@@ -0,0 +1,18 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Car Parts,Mamba4Cast,MASE,1.061,
+accuracy,CIF 2016,Mamba4Cast,MASE,0.925,
+accuracy,Covid Deaths,Mamba4Cast,MASE,5.926,
+accuracy,ERCOT Load,Mamba4Cast,MASE,0.657,
+accuracy,Exchange Rate,Mamba4Cast,MASE,1.329,
+accuracy,FRED-MD,Mamba4Cast,MASE,0.524,
+accuracy,Hospital,Mamba4Cast,MASE,0.806,
+accuracy,M1 (Monthly),Mamba4Cast,MASE,1.100,
+accuracy,M1 (Quarterly),Mamba4Cast,MASE,1.695,
+accuracy,M3 (Monthly),Mamba4Cast,MASE,0.849,
+accuracy,M3 (Quarterly),Mamba4Cast,MASE,1.251,
+accuracy,NN5 (Daily),Mamba4Cast,MASE,0.833,
+accuracy,NN5 (Weekly),Mamba4Cast,MASE,0.956,
+accuracy,Tourism (Monthly),Mamba4Cast,MASE,1.567,
+accuracy,Tourism (Quarterly),Mamba4Cast,MASE,1.746,
+accuracy,Traffic,Mamba4Cast,MASE,1.120,
+accuracy,Weather,Mamba4Cast,MASE,0.726,
diff --git a/result/per_paper/2410.09385/accuracy_efficiency_traced.csv b/result/per_paper/2410.09385/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..3eaab66b520f2a84fca88a241316d05b1e643c9d
--- /dev/null
+++ b/result/per_paper/2410.09385/accuracy_efficiency_traced.csv
@@ -0,0 +1,18 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Car Parts,Mamba4Cast,MASE,1.061,,0,2,1
+accuracy,CIF 2016,Mamba4Cast,MASE,0.925,,0,3,1
+accuracy,Covid Deaths,Mamba4Cast,MASE,5.926,,0,4,1
+accuracy,ERCOT Load,Mamba4Cast,MASE,0.657,,0,5,1
+accuracy,Exchange Rate,Mamba4Cast,MASE,1.329,,0,6,1
+accuracy,FRED-MD,Mamba4Cast,MASE,0.524,,0,7,1
+accuracy,Hospital,Mamba4Cast,MASE,0.806,,0,8,1
+accuracy,M1 (Monthly),Mamba4Cast,MASE,1.100,,0,9,1
+accuracy,M1 (Quarterly),Mamba4Cast,MASE,1.695,,0,10,1
+accuracy,M3 (Monthly),Mamba4Cast,MASE,0.849,,0,11,1
+accuracy,M3 (Quarterly),Mamba4Cast,MASE,1.251,,0,12,1
+accuracy,NN5 (Daily),Mamba4Cast,MASE,0.833,,0,13,1
+accuracy,NN5 (Weekly),Mamba4Cast,MASE,0.956,,0,14,1
+accuracy,Tourism (Monthly),Mamba4Cast,MASE,1.567,,0,15,1
+accuracy,Tourism (Quarterly),Mamba4Cast,MASE,1.746,,0,16,1
+accuracy,Traffic,Mamba4Cast,MASE,1.120,,0,17,1
+accuracy,Weather,Mamba4Cast,MASE,0.726,,0,18,1
diff --git a/result/per_paper/2410.09385/components_architecture.csv b/result/per_paper/2410.09385/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a76312d092b3ac805bfbe465b03c4cacee9137fd
--- /dev/null
+++ b/result/per_paper/2410.09385/components_architecture.csv
@@ -0,0 +1,10 @@
+component,what_it_is,provenance,citation,evidence
+Pre-processing,Scales input time series using a Min-Max Scaler and extracts time features for positional embeddings.,proposed_here,,we scale the input series using a Min-Max Scaler and extract time features for positional embeddings.
+Input Embedding,Embeds scaled input values and temporal information using convolutions with different dilations to ensure a large receptive field.,proposed_here,,"we embed the scaled input values and their temporal information using convolutions with different dilations, ensuring a large receptive field for the representation used by future layers."
+Value Linear Embedding,Linear transformation applied to input token embeddings.,reused_cited,"Hollmann et al., 2023; Dooley et al., 2023","D[""Value Linear Embedding""] --> E[""concat""]"
+Position Linear Embedding,Linear transformation applied to positional information for temporal context.,reused_cited,"Hollmann et al., 2023; Dooley et al., 2023","F[""Position Linear Embedding""] --> E[""concat""]"
+Convolutional Block (Conv+ConvId+GeLU),Stacked convolutional layers with identity mapping and GeLU activation for feature extraction.,reused_cited,"Katharopoulos et al., 2020; Yang et al., 2024a,b","E --> G[""Stacked Conv+ ConvId + GeLU""]"
+Encoder Block,Comprises Mamba2 blocks with LayerNorm and additional dilated convolution layers for encoding temporal dependencies.,proposed_here,,Encoder: comprises of Mamba2 blocks with LayerNorm to avoid noisy learning signals followed by another dilated convolution layer.
+Mamba2 Block,Linear recurrent neural network with scalar multiple of identity matrix for efficient computation.,reused_cited,"Dao & Gu, 2024",Mamba2 is a linear Recurrent Neural Network described by the following recurrence...
+LayerNorm,Normalization applied to stabilize training and improve gradient flow.,reused_cited,"Ba et al., 2016","H[""Stacked Conv+ GeLU""] --> I[""LayerNorm""]"
+Decoder,Linear projection layer that transforms embedded token representations into point forecasts.,proposed_here,,Decoder: the final component is a linear projection layer that transforms the embedded token representations into point forecasts.
diff --git a/result/per_paper/2410.09385/computational.csv b/result/per_paper/2410.09385/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..1ca1e82a82f1598382b263a6f6fef7a52f85c525
--- /dev/null
+++ b/result/per_paper/2410.09385/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,Nvidia RTX2080Ti,,stated,Training was conducted over 3 days on a single Nvidia RTX2080Ti GPU
+num_devices,1,,stated,Training was conducted over 3 days on a single Nvidia RTX2080Ti GPU
+training_cost,,,not_reported,
+training_batch_size,64,,stated,trained for 420K batches of size 64
+training_steps_or_epochs,420000,,stated,trained for 420K batches of size 64
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,512,,stated,evaluate on ... with a 512 context length
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2410.10393/accuracy_efficiency.csv b/result/per_paper/2410.10393/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2410.10393/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2410.10393/accuracy_efficiency_traced.csv b/result/per_paper/2410.10393/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2410.10393/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2410.10393/components_architecture.csv b/result/per_paper/2410.10393/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..fec728ddabea8ab88d5765fac35b6dab1eff9b5e
--- /dev/null
+++ b/result/per_paper/2410.10393/components_architecture.csv
@@ -0,0 +1,10 @@
+component,what_it_is,provenance,citation,evidence
+Pretraining Dataset,A non-leaking pretraining dataset containing approximately 230 billion data points to facilitate effective pretraining of foundation models.,proposed_here,,"To facilitate the effective pretraining and evaluation of foundation models, we also provide a non-leaking pretraining dataset containing approximately 230 billion data points."
+Train/Test Datasets,"A collection of 23 datasets spanning 144,000 time series and 177 million data points across seven domains, 10 frequencies, multivariate inputs, and prediction lengths ranging from short to long-term forecasts.",proposed_here,,"GIFT-Eval encompasses 23 datasets over 144,000 time series and 177 million data points, spanning seven domains, 10 frequencies, multivariate inputs, and prediction lengths ranging from short to long-term forecasts."
+Domain Coverage,Seven distinct domains represented in the benchmark to ensure diversity in time series characteristics.,proposed_here,,"GIFT-Eval encompasses 23 datasets [...] spanning seven domains, 10 frequencies, multivariate inputs, and prediction lengths ranging from short to long-term forecasts."
+Frequency Coverage,Ten different time series frequencies included to evaluate models across varying temporal granularities.,proposed_here,,"GIFT-Eval encompasses 23 datasets [...] spanning seven domains, 10 frequencies, multivariate inputs, and prediction lengths ranging from short to long-term forecasts."
+Prediction Lengths,A range of prediction lengths from short-term to long-term forecasts to assess model adaptability.,proposed_here,,GIFT-Eval encompasses 23 datasets [...] prediction lengths ranging from short to long-term forecasts.
+Multivariate Inputs,"Support for multivariate time series inputs to evaluate models on complex, real-world scenarios.",proposed_here,,"GIFT-Eval encompasses 23 datasets [...] multivariate inputs, and prediction lengths ranging from short to long-term forecasts."
+Baseline Models,"A comprehensive analysis of 17 baselines, including statistical, deep learning, and foundation models, evaluated on GIFT-Eval.",proposed_here,,"Additionally, we provide a comprehensive analysis of 17 baselines, which includes statistical models, deep learning models, and foundation models."
+Qualitative Analysis,"A detailed qualitative analysis of model performance, including failure cases of both deep learning and foundation models.",proposed_here,,We further provide a qualitative analysis showing failure cases of both deep learning and foundation models.
+Leaderboard,A public leaderboard to track model performance on GIFT-Eval.,proposed_here,,"Code, data, and the leaderboard can be found at https://github.com/SalesforceAIResearch/gift-eval."
diff --git a/result/per_paper/2410.10393/computational.csv b/result/per_paper/2410.10393/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a9f784acde890d4314ed74715b321cd3d0f24099
--- /dev/null
+++ b/result/per_paper/2410.10393/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No information on hardware (e.g., GPU/TPU) used for training or inference."
+num_devices,,,not_reported,"No mention of the number of devices (e.g., GPUs) used."
+training_cost,,,not_reported,No financial or resource cost details for training.
+training_batch_size,,,not_reported,No batch size information for training.
+training_steps_or_epochs,,,not_reported,No details on training steps or epochs.
+precision,,,not_reported,"No mention of precision (e.g., FP16/FP32)."
+inference_latency,,,not_reported,No latency measurements for inference.
+inference_throughput,,,not_reported,No throughput metrics for inference.
+peak_memory,,,not_reported,No memory usage details reported.
+flops_or_macs,,,not_reported,No computational complexity (FLOPs/MACs) mentioned.
+num_inference_samples,,,not_reported,No explicit count of inference samples.
+params,,,not_reported,Model parameter count not provided.
+context_lengths_evaluated,,,not_reported,Context length (l) is defined theoretically but not quantified.
+horizon_lengths_evaluated,,,not_reported,Forecast horizon (h) is discussed categorically (short/medium/long) but not numerically.
+inference_batch_size,,,not_reported,No batch size details for inference.
diff --git a/result/per_paper/2410.10469/accuracy_efficiency.csv b/result/per_paper/2410.10469/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d886dc32f17694083380b6d046475a95c118fe1d
--- /dev/null
+++ b/result/per_paper/2410.10469/accuracy_efficiency.csv
@@ -0,0 +1,61 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,M1 Monthly,Moirai-MoE-Small,MAE,1,
+accuracy,M1 Monthly,Moirai-MoE-Base,MAE,1,
+accuracy,M3 Monthly,Moirai-MoE-Small,MAE,646.07,
+accuracy,M3 Monthly,Moirai-MoE-Base,MAE,617.31,
+accuracy,M3 Other,Moirai-MoE-Small,MAE,185.89,
+accuracy,M3 Other,Moirai-MoE-Base,MAE,179.92,
+accuracy,M4 Monthly,Moirai-MoE-Small,MAE,569.25,
+accuracy,M4 Monthly,Moirai-MoE-Base,MAE,544.08,
+accuracy,M4 Weekly,Moirai-MoE-Small,MAE,302.65,
+accuracy,M4 Weekly,Moirai-MoE-Base,MAE,278.37,
+accuracy,M4 Daily,Moirai-MoE-Small,MAE,172.45,
+accuracy,M4 Daily,Moirai-MoE-Base,MAE,163.40,
+accuracy,M4 Hourly,Moirai-MoE-Small,MAE,241.58,
+accuracy,M4 Hourly,Moirai-MoE-Base,MAE,217.35,
+accuracy,Tourism Quarterly,Moirai-MoE-Small,MAE,9,
+accuracy,Tourism Quarterly,Moirai-MoE-Base,MAE,7,
+accuracy,Tourism Monthly,Moirai-MoE-Small,MAE,2,
+accuracy,Tourism Monthly,Moirai-MoE-Base,MAE,2,
+accuracy,CIF 2016,Moirai-MoE-Small,MAE,453,
+accuracy,CIF 2016,Moirai-MoE-Base,MAE,568,
+accuracy,Aus. Elec. Demand,Moirai-MoE-Small,MAE,215.28,
+accuracy,Aus. Elec. Demand,Moirai-MoE-Base,MAE,227.92,
+accuracy,Bitcoin,Moirai-MoE-Small,MAE,1.55,
+accuracy,Bitcoin,Moirai-MoE-Base,MAE,1.90,
+accuracy,Pedestrian Counts,Moirai-MoE-Small,MAE,41.35,
+accuracy,Pedestrian Counts,Moirai-MoE-Base,MAE,32.37,
+accuracy,Vehicle Trips,Moirai-MoE-Small,MAE,21.62,
+accuracy,Vehicle Trips,Moirai-MoE-Base,MAE,21.65,
+accuracy,KDD Cup 2018,Moirai-MoE-Small,MAE,40.21,
+accuracy,KDD Cup 2018,Moirai-MoE-Base,MAE,40.86,
+accuracy,Australia Weather,Moirai-MoE-Small,MAE,1.76,
+accuracy,Australia Weather,Moirai-MoE-Base,MAE,1.75,
+accuracy,NN5 Daily,Moirai-MoE-Small,MAE,4.04,
+accuracy,NN5 Daily,Moirai-MoE-Base,MAE,3.49,
+accuracy,NN5 Weekly,Moirai-MoE-Small,MAE,15.74,
+accuracy,NN5 Weekly,Moirai-MoE-Base,MAE,15.29,
+accuracy,Carparts,Moirai-MoE-Small,MAE,0.45,
+accuracy,Carparts,Moirai-MoE-Base,MAE,0.44,
+accuracy,FRED-MD,Moirai-MoE-Small,MAE,1,
+accuracy,FRED-MD,Moirai-MoE-Base,MAE,2,
+accuracy,Traffic Hourly,Moirai-MoE-Small,MAE,0.013,
+accuracy,Traffic Hourly,Moirai-MoE-Base,MAE,0.014,
+accuracy,Traffic Weekly,Moirai-MoE-Small,MAE,1.13,
+accuracy,Traffic Weekly,Moirai-MoE-Base,MAE,1.14,
+accuracy,Rideshare,Moirai-MoE-Small,MAE,1.26,
+accuracy,Rideshare,Moirai-MoE-Base,MAE,1.26,
+accuracy,Hospital,Moirai-MoE-Small,MAE,20.17,
+accuracy,Hospital,Moirai-MoE-Base,MAE,19.60,
+accuracy,COVID Deaths,Moirai-MoE-Small,MAE,119.00,
+accuracy,COVID Deaths,Moirai-MoE-Base,MAE,102.92,
+accuracy,Temperature Rain,Moirai-MoE-Small,MAE,5.33,
+accuracy,Temperature Rain,Moirai-MoE-Base,MAE,5.36,
+accuracy,Sunspot,Moirai-MoE-Small,MAE,0.10,
+accuracy,Sunspot,Moirai-MoE-Base,MAE,0.08,
+accuracy,Saugeen River Flow,Moirai-MoE-Small,MAE,23.05,
+accuracy,Saugeen River Flow,Moirai-MoE-Base,MAE,24.40,
+accuracy,US Births,Moirai-MoE-Small,MAE,411.61,
+accuracy,US Births,Moirai-MoE-Base,MAE,385.24,
+efficiency,Spent Time (s),Moirai-MoE-S,Spent Time (s),273,
+efficiency,Spent Time (s),Moirai-MoE-B,Spent Time (s),370,
diff --git a/result/per_paper/2410.10469/accuracy_efficiency_traced.csv b/result/per_paper/2410.10469/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5c501c767fa7cc75a5fb49c73529ad9ccd2ba672
--- /dev/null
+++ b/result/per_paper/2410.10469/accuracy_efficiency_traced.csv
@@ -0,0 +1,61 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,M1 Monthly,Moirai-MoE-Small,MAE,1,,0,1,10
+accuracy,M1 Monthly,Moirai-MoE-Base,MAE,1,,0,1,11
+accuracy,M3 Monthly,Moirai-MoE-Small,MAE,646.07,,0,2,10
+accuracy,M3 Monthly,Moirai-MoE-Base,MAE,617.31,,0,2,11
+accuracy,M3 Other,Moirai-MoE-Small,MAE,185.89,,0,3,10
+accuracy,M3 Other,Moirai-MoE-Base,MAE,179.92,,0,3,11
+accuracy,M4 Monthly,Moirai-MoE-Small,MAE,569.25,,0,4,10
+accuracy,M4 Monthly,Moirai-MoE-Base,MAE,544.08,,0,4,11
+accuracy,M4 Weekly,Moirai-MoE-Small,MAE,302.65,,0,5,10
+accuracy,M4 Weekly,Moirai-MoE-Base,MAE,278.37,,0,5,11
+accuracy,M4 Daily,Moirai-MoE-Small,MAE,172.45,,0,6,10
+accuracy,M4 Daily,Moirai-MoE-Base,MAE,163.40,,0,6,11
+accuracy,M4 Hourly,Moirai-MoE-Small,MAE,241.58,,0,7,10
+accuracy,M4 Hourly,Moirai-MoE-Base,MAE,217.35,,0,7,11
+accuracy,Tourism Quarterly,Moirai-MoE-Small,MAE,9,,0,8,10
+accuracy,Tourism Quarterly,Moirai-MoE-Base,MAE,7,,0,8,11
+accuracy,Tourism Monthly,Moirai-MoE-Small,MAE,2,,0,9,10
+accuracy,Tourism Monthly,Moirai-MoE-Base,MAE,2,,0,9,11
+accuracy,CIF 2016,Moirai-MoE-Small,MAE,453,,0,10,10
+accuracy,CIF 2016,Moirai-MoE-Base,MAE,568,,0,10,11
+accuracy,Aus. Elec. Demand,Moirai-MoE-Small,MAE,215.28,,0,11,10
+accuracy,Aus. Elec. Demand,Moirai-MoE-Base,MAE,227.92,,0,11,11
+accuracy,Bitcoin,Moirai-MoE-Small,MAE,1.55,,0,12,10
+accuracy,Bitcoin,Moirai-MoE-Base,MAE,1.90,,0,12,11
+accuracy,Pedestrian Counts,Moirai-MoE-Small,MAE,41.35,,0,13,10
+accuracy,Pedestrian Counts,Moirai-MoE-Base,MAE,32.37,,0,13,11
+accuracy,Vehicle Trips,Moirai-MoE-Small,MAE,21.62,,0,14,10
+accuracy,Vehicle Trips,Moirai-MoE-Base,MAE,21.65,,0,14,11
+accuracy,KDD Cup 2018,Moirai-MoE-Small,MAE,40.21,,0,15,10
+accuracy,KDD Cup 2018,Moirai-MoE-Base,MAE,40.86,,0,15,11
+accuracy,Australia Weather,Moirai-MoE-Small,MAE,1.76,,0,16,10
+accuracy,Australia Weather,Moirai-MoE-Base,MAE,1.75,,0,16,11
+accuracy,NN5 Daily,Moirai-MoE-Small,MAE,4.04,,0,17,10
+accuracy,NN5 Daily,Moirai-MoE-Base,MAE,3.49,,0,17,11
+accuracy,NN5 Weekly,Moirai-MoE-Small,MAE,15.74,,0,18,10
+accuracy,NN5 Weekly,Moirai-MoE-Base,MAE,15.29,,0,18,11
+accuracy,Carparts,Moirai-MoE-Small,MAE,0.45,,0,19,10
+accuracy,Carparts,Moirai-MoE-Base,MAE,0.44,,0,19,11
+accuracy,FRED-MD,Moirai-MoE-Small,MAE,1,,0,20,10
+accuracy,FRED-MD,Moirai-MoE-Base,MAE,2,,0,20,11
+accuracy,Traffic Hourly,Moirai-MoE-Small,MAE,0.013,,0,21,10
+accuracy,Traffic Hourly,Moirai-MoE-Base,MAE,0.014,,0,21,11
+accuracy,Traffic Weekly,Moirai-MoE-Small,MAE,1.13,,0,22,10
+accuracy,Traffic Weekly,Moirai-MoE-Base,MAE,1.14,,0,22,11
+accuracy,Rideshare,Moirai-MoE-Small,MAE,1.26,,0,23,10
+accuracy,Rideshare,Moirai-MoE-Base,MAE,1.26,,0,23,11
+accuracy,Hospital,Moirai-MoE-Small,MAE,20.17,,0,24,10
+accuracy,Hospital,Moirai-MoE-Base,MAE,19.60,,0,24,11
+accuracy,COVID Deaths,Moirai-MoE-Small,MAE,119.00,,0,25,10
+accuracy,COVID Deaths,Moirai-MoE-Base,MAE,102.92,,0,25,11
+accuracy,Temperature Rain,Moirai-MoE-Small,MAE,5.33,,0,26,10
+accuracy,Temperature Rain,Moirai-MoE-Base,MAE,5.36,,0,26,11
+accuracy,Sunspot,Moirai-MoE-Small,MAE,0.10,,0,27,10
+accuracy,Sunspot,Moirai-MoE-Base,MAE,0.08,,0,27,11
+accuracy,Saugeen River Flow,Moirai-MoE-Small,MAE,23.05,,0,28,10
+accuracy,Saugeen River Flow,Moirai-MoE-Base,MAE,24.40,,0,28,11
+accuracy,US Births,Moirai-MoE-Small,MAE,411.61,,0,29,10
+accuracy,US Births,Moirai-MoE-Base,MAE,385.24,,0,29,11
+efficiency,Spent Time (s),Moirai-MoE-S,Spent Time (s),273,,3,1,7
+efficiency,Spent Time (s),Moirai-MoE-B,Spent Time (s),370,,3,1,8
diff --git a/result/per_paper/2410.10469/components_architecture.csv b/result/per_paper/2410.10469/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5dd2e96f0215ec9256f3332e6c58e7d183e34241
--- /dev/null
+++ b/result/per_paper/2410.10469/components_architecture.csv
@@ -0,0 +1,4 @@
+component,what_it_is,provenance,citation,evidence
+Input/Output Projection Layer,"A single projection layer used for both input and output processing, replacing frequency-specific layers in prior models.",proposed_here,,MOIRAI-MOE uses a single input/output projection layer while delegating the modeling of diverse time series patterns to the sparse mixture of experts (MoE) within Transformers.
+Sparse Mixture of Experts (MoE),A mechanism within Transformer layers that dynamically routes tokens to specialized expert subnetworks for processing.,reused_cited,"Lepikhin et al., 2021; Fedus et al., 2022; Dai et al., 2024",The core idea of MOIRAI-MOE is to utilize a single input/output projection layer while delegating the modeling of diverse time series patterns to the sparse specialized experts in Transformer layers.
+Expert Gating Function,A function that determines which expert subnetworks are activated for a given token based on its features.,reused_cited,"Lepikhin et al., 2021; Fedus et al., 2022; Dai et al., 2024",This study investigates existing expert gating functions that generally use a [method] to route tokens to experts.
diff --git a/result/per_paper/2410.10469/computational.csv b/result/per_paper/2410.10469/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a170c68943fdb7b782a5ba84513a28574f361fb1
--- /dev/null
+++ b/result/per_paper/2410.10469/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,A100 (40G) GPUs,,stated,All MOIRAI-MOE models are trained on 16 A100 (40G) GPUs...
+num_devices,16,,stated,All MOIRAI-MOE models are trained on 16 A100 (40G) GPUs...
+training_cost,,,not_reported,
+training_batch_size,1024,,stated,"trained on 16 A100 (40G) GPUs using a batch size of 1,024..."
+training_steps_or_epochs,,,not_reported,
+precision,bfloat16,,stated,"using a batch size of 1,024 and bfloat16 precision."
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,20,,stated,"number of sampling samples to 20, aligning with the settings used in Chronos."
+params,117,M,stated,total parameter sizes of 117M for MOIRAI-MOES...
+context_lengths_evaluated,512,,stated,set the context length to 512...
+horizon_lengths_evaluated,16,time steps,stated,predicted token is one (corresponding to 16 time steps).
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2410.11802/accuracy_efficiency.csv b/result/per_paper/2410.11802/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2410.11802/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2410.11802/accuracy_efficiency_traced.csv b/result/per_paper/2410.11802/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2410.11802/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2410.11802/components_architecture.csv b/result/per_paper/2410.11802/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f6e4a7be8d0d87a001469e85b37eace05a83da23
--- /dev/null
+++ b/result/per_paper/2410.11802/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Time Series Foundation Models (TSFMs),"Pre-trained models on massive heterogeneous time series data, designed to generalize across domains and tasks in time series forecasting.",proposed_here,,Time Series Foundation Models (TSFMs) that are pre-trained on massive heterogeneous time series data aim to overcome these limitations.
+Patching Operations,"A technique to capture temporal dependencies at the patch level, introduced by Triformer and further utilized in PatchTST.",reused_cited,"Triformer [7], PatchTST [45]","patching operations, initially introduced by Triformer [7], significantly enhance models' capabilities to capture temporal dependencies at the patch level. And PatchTST [45] lies in establishing the fundamental role of patching operations in TSF tasks."
+Large Language Models (LLMs),"A class of models used as a basis for some TSFMs, leveraging representation learning capabilities for time series analysis.",reused_cited,"Cited in the paper (e.g., [24])",one study [24] proposes a framework for LLM-based time series analysis.
+Pre-training on Heterogeneous Time Series Data,"A training paradigm for TSFMs involving diverse, domain-agnostic time series data to enable generalizability.",proposed_here,,TSFMs that are pre-trained on massive heterogeneous time series data aim to overcome these limitations.
+"Model Architectures (e.g., Transformer-based)","Architectural designs of TSFMs, including transformer-based structures, adapted for time series forecasting tasks.",proposed_here,,"TSFM-Bench covers a wide range of TSFMs, including those based on large language models and those pre-trained on time series data."
diff --git a/result/per_paper/2410.11802/computational.csv b/result/per_paper/2410.11802/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2410.11802/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2411.02941/accuracy_efficiency.csv b/result/per_paper/2411.02941/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..86e606f5c8f1163a3c6a5b823faeda0e4dcbee15
--- /dev/null
+++ b/result/per_paper/2411.02941/accuracy_efficiency.csv
@@ -0,0 +1,5 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTm2,TSMamba,MSE and MAE,0.274,
+accuracy,AVG,TSMamba,MSE and MAE,0.328,
+accuracy,Weather,TSMamba,MSE and MAE,0.211,
+accuracy,Average,TSMamba,MSE and MAE,0.328,
diff --git a/result/per_paper/2411.02941/accuracy_efficiency_traced.csv b/result/per_paper/2411.02941/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6fd6a9b5a509f9ee84142848850cfc8e751af93e
--- /dev/null
+++ b/result/per_paper/2411.02941/accuracy_efficiency_traced.csv
@@ -0,0 +1,5 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTm2,TSMamba,MSE and MAE,0.274,,1,3,13
+accuracy,AVG,TSMamba,MSE and MAE,0.328,,1,7,13
+accuracy,Weather,TSMamba,MSE and MAE,0.211,,1,8,13
+accuracy,Average,TSMamba,MSE and MAE,0.328,,1,13,13
diff --git a/result/per_paper/2411.02941/components_architecture.csv b/result/per_paper/2411.02941/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..bbadefe0c820fbeae6aa84c30f82bf692b05b148
--- /dev/null
+++ b/result/per_paper/2411.02941/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Forward Mamba Encoder,A Mamba-based sequence model that processes time series data in the forward direction to capture temporal dependencies with linear complexity.,reused_cited,"Gu & Dao, 2023","TSMamba is built on the Mamba architecture, which enhances structured state space sequence models (SSMs) by making parameters functions of the inputs, allowing selective information propagation."
+Backward Mamba Encoder,A Mamba-based sequence model that processes time series data in the backward direction to capture temporal dependencies with linear complexity.,reused_cited,"Gu & Dao, 2023","TSMamba combines forward and backward Mamba encoders to capture temporal dependencies with linear complexity, achieving high predictive accuracy."
+Two-Stage Transfer Learning Process,"A training methodology that first optimizes the forward/backward encoders via patch-wise autoregressive prediction, then trains a prediction head and refines components for long-term forecasting.",proposed_here,,"In the first stage, the forward and backward backbones are optimized via patch-wise autoregressive prediction; in the second stage, the model trains a prediction head and refines other components for long-term forecasting."
+Channel-Wise Compressed Attention Module,"A module introduced during fine-tuning to capture cross-channel dependencies in multivariate datasets, addressing the assumption of channel independence in the pretrained model.",proposed_here,,A channel-wise compressed attention module is introduced to capture cross-channel dependencies during fine-tuning on specific multivariate datasets.
+Prediction Head,A task-specific output layer trained in the second stage of the transfer learning process to generate forecasts.,proposed_here,,The model trains a prediction head and refines other components for long-term forecasting.
+Channel Independence (CI) Assumption,A design choice in the pretrained Mamba backbone to handle varying channel numbers across datasets by assuming independence between channels.,proposed_here,,The pretrained model assumes channel independence (CI) to handle varying channel numbers across datasets.
diff --git a/result/per_paper/2411.02941/computational.csv b/result/per_paper/2411.02941/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..9ebe71861f48a3e4a6e9e113b6fd811127dc56eb
--- /dev/null
+++ b/result/per_paper/2411.02941/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,"[96, 192, 336, 720]",steps,stated,The paper defines T as the prediction length (target window) and evaluates TSMamba on these values in the tables.
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2501.02945/accuracy_efficiency.csv b/result/per_paper/2501.02945/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2501.02945/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2501.02945/accuracy_efficiency_traced.csv b/result/per_paper/2501.02945/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2501.02945/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2501.02945/components_architecture.csv b/result/per_paper/2501.02945/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..7bd53a49d33df2603ef2151273acccab8952c3a3
--- /dev/null
+++ b/result/per_paper/2501.02945/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Temporal Featurization Module,"A lightweight feature encoding scheme that captures time progression, multi-scale seasonality, and covariates to represent time series data as tabular inputs for TabPFN-v2.",proposed_here,,"We design a compact feature scheme that encodes time progression, multi-scale seasonality, and covariates, allowing tabular models to act on temporal data effectively."
+TabPFN-v2 Pretrained Model,"A tabular foundation model pretrained on diverse supervised tasks (classification, regression, etc.), reused as the core regressor for time series forecasting in TabPFN-TS.",reused_cited,"Hollmann et al., 2025","TabPFN-v2 (Hollmann et al., 2025) is notable as the first tabular foundation model, supporting many tasks, including classification, regression, outlier detection, density estimation, synthetic data generation, embeddings that are useful for downstream tasks, and fine-tunability."
+Tabular Regression Head,"The output layer of TabPFN-v2 adapted to predict future time series values as a regression task, using temporal features and covariates as inputs.",proposed_here,,Forecasting then reduces to predicting future rows—whose temporal features are known in advance—using TabPFN-v2 as a tabular regressor.
+Time Index Encoding,"A feature that explicitly encodes the position of each time step (e.g., 0, 1, 2, ...) to provide temporal reference within the timeline.",proposed_here,,"To introduce a temporal reference within the timeline, we include the index of each time step as a feature (e.g., 0 for the first time step in the time series, 4 for..."
+Covariate Integration Layer,"A mechanism to incorporate external covariates (e.g., holidays, prices) alongside temporal features for covariate-informed forecasting.",proposed_here,,"The goal is to predict future observations [...] optionally, a set of covariates—external variables that provide additional information about the system's dynamics."
diff --git a/result/per_paper/2501.02945/computational.csv b/result/per_paper/2501.02945/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d91ba7f3a42cd64cff8bf5729085421faa470c20
--- /dev/null
+++ b/result/per_paper/2501.02945/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA T4,,stated,"TabPFN-TS runs across four T4s, and we report time normalized to one GPU."
+num_devices,4,,stated,TabPFN-TS runs across four T4s.
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,4096,points,stated,Figure A.8: Effect of context length on forecasting performance of TabPFN-TS. ... gains diminish beyond 4096 points.
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2501.05675/accuracy_efficiency.csv b/result/per_paper/2501.05675/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2501.05675/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2501.05675/accuracy_efficiency_traced.csv b/result/per_paper/2501.05675/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2501.05675/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2501.05675/components_architecture.csv b/result/per_paper/2501.05675/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..fa65e4f165c1de13400127ed572a40fadb473868
--- /dev/null
+++ b/result/per_paper/2501.05675/components_architecture.csv
@@ -0,0 +1,3 @@
+component,what_it_is,provenance,citation,evidence
+Model Alignment Module,A component designed to address the misalignment between the expression domains of LLMs and task-specific models by aligning their anomaly score interpretations.,proposed_here,"Chen et al., 2023","The LLM and the TSADM interpret anomaly scores differently, meaning they may use the same score to represent different levels of anomaly severity. [...] Such inconsistencies in score interpretation can disrupt effective collaboration between the two models."
+Collaborative Loss Function,A loss function introduced to mitigate prediction error accumulation during collaboration between LLMs and task-specific models by ensuring errors do not compound.,proposed_here,"Chen et al., 2023","Both the LLM and the TSADM are subject to prediction errors. [...] errors from the two models tend to either compound or settle at a compromise between the higher and lower values, rather than canceling each other."
diff --git a/result/per_paper/2501.05675/computational.csv b/result/per_paper/2501.05675/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6b4abe392dd38d09b021099491656975bc12e0fa
--- /dev/null
+++ b/result/per_paper/2501.05675/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware specifications (e.g., GPU/TPU types) are mentioned for the model."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not reported."
+training_cost,,,not_reported,"Training cost (e.g., monetary cost, computational resources) is not mentioned."
+training_batch_size,,,not_reported,Training batch size is not specified in the text.
+training_steps_or_epochs,,,not_reported,Training steps or epochs are not reported.
+precision,,,not_reported,"Precision refers to evaluation metrics (e.g., 0.82–1.0 in the table), not computational precision (e.g., 32-bit vs. 16-bit)."
+inference_latency,,,not_reported,"Inference latency (e.g., time per sample) is not mentioned."
+inference_throughput,,,not_reported,"Inference throughput (e.g., samples per second) is not reported."
+peak_memory,,,not_reported,"Peak memory usage (e.g., VRAM) is not specified."
+flops_or_macs,,,not_reported,FLOPs or MACs (computational complexity) are not mentioned.
+num_inference_samples,,,not_reported,The number of inference samples tested is not explicitly stated.
+params,,,not_reported,Model parameter count is not reported.
+context_lengths_evaluated,,,not_reported,"Context lengths (e.g., input sequence lengths) evaluated are not specified."
+horizon_lengths_evaluated,,,not_reported,"Horizon lengths (e.g., prediction window) evaluated are not mentioned."
+inference_batch_size,,,not_reported,Inference batch size is not reported.
diff --git a/result/per_paper/2501.14170/accuracy_efficiency.csv b/result/per_paper/2501.14170/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..65795afed1f1d3e2ca4a0bc09cac0c192ca3c014
--- /dev/null
+++ b/result/per_paper/2501.14170/accuracy_efficiency.csv
@@ -0,0 +1,11 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,KPI-02e99,Argos w/o Aggregator,F1 Score,0.96,
+accuracy,KPI-02e99,Argos,F1 Score,0.99,
+accuracy,KPI-07927,Argos w/o Aggregator,F1 Score,0.67,
+accuracy,KPI-07927,Argos,F1 Score,0.99,
+accuracy,KPI-1c35d,Argos w/o Aggregator,F1 Score,0.80,
+accuracy,KPI-1c35d,Argos,F1 Score,0.99,
+accuracy,Yahoo-A2,Argos w/o Aggregator,F1 Score,0.87,
+accuracy,Yahoo-A2,Argos,F1 Score,0.95,
+accuracy,Yahoo-A4,Argos w/o Aggregator,F1 Score,0.84,
+accuracy,Yahoo-A4,Argos,F1 Score,0.87,
diff --git a/result/per_paper/2501.14170/accuracy_efficiency_traced.csv b/result/per_paper/2501.14170/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..659469bb3dfc3efcba516665aba06e439a76bd90
--- /dev/null
+++ b/result/per_paper/2501.14170/accuracy_efficiency_traced.csv
@@ -0,0 +1,11 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,KPI-02e99,Argos w/o Aggregator,F1 Score,0.96,,1,1,2
+accuracy,KPI-02e99,Argos,F1 Score,0.99,,1,1,3
+accuracy,KPI-07927,Argos w/o Aggregator,F1 Score,0.67,,1,2,2
+accuracy,KPI-07927,Argos,F1 Score,0.99,,1,2,3
+accuracy,KPI-1c35d,Argos w/o Aggregator,F1 Score,0.80,,1,3,2
+accuracy,KPI-1c35d,Argos,F1 Score,0.99,,1,3,3
+accuracy,Yahoo-A2,Argos w/o Aggregator,F1 Score,0.87,,1,4,2
+accuracy,Yahoo-A2,Argos,F1 Score,0.95,,1,4,3
+accuracy,Yahoo-A4,Argos w/o Aggregator,F1 Score,0.84,,1,5,2
+accuracy,Yahoo-A4,Argos,F1 Score,0.87,,1,5,3
diff --git a/result/per_paper/2501.14170/components_architecture.csv b/result/per_paper/2501.14170/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4101d7a00c747f846c28bcc264161805e0e1789d
--- /dev/null
+++ b/result/per_paper/2501.14170/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Agent-based Pipeline with Feedback Loops,"A collaborative framework where multiple agents iteratively propose, validate, fix, and refine anomaly detection rules to reduce syntax errors and improve accuracy.",proposed_here,,ARGOS employs an agent-based pipeline with feedback loops to iteratively correct anomaly detection rules and improve accuracy.
+Anomaly Rules (LLM-Generated),Explainable and reproducible anomaly detection rules autonomously generated by large language models (LLMs) as an intermediate representation for time-series analysis.,proposed_here,,ARGOS proposes to use explainable and reproducible anomaly rules as intermediate representation and employs LLMs to autonomously generate such rules.
+Aggregation Algorithm for Prediction Merging,A method to combine predictions from LLM-generated rules and existing anomaly detection systems during runtime inference to produce final anomaly predictions.,proposed_here,,ARGOS uses an aggregation algorithm to merge the predictions from both the rules and the existing anomaly detectors to generate the final anomaly prediction.
+Rule Proposal Mechanism (n Candidates per Iteration),A process that simultaneously proposes a set of n rule candidates in each iteration and selects the best k rules for refinement in subsequent iterations.,proposed_here,,ARGOS simultaneously proposes a set of n rule candidates in each iteration and selects the best k rules for further refinement in the next iteration.
+Existing Anomaly Detector (Integration),"A pre-existing anomaly detection system used during training to identify incorrect samples, which ARGOS leverages to learn data patterns.",reused_cited,Cited in the paper's context of integrating with existing systems,"During training, ARGOS primarily learns data patterns from the incorrect samples identified by an existing anomaly detector."
+Deployment Module for Online Anomaly Detection,"A component responsible for deploying trained anomaly rules for low-cost, real-time monitoring of cloud infrastructure metrics.",proposed_here,,The system will [...] deploy the trained rules for low-cost online anomaly detection.
diff --git a/result/per_paper/2501.14170/computational.csv b/result/per_paper/2501.14170/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d101922e7599eb8c4800ed3c7930e4ffc3614bf7
--- /dev/null
+++ b/result/per_paper/2501.14170/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,A100,GPU,stated,Figure 1: GPU utilization and memory usage metrics for a distributed model training on 256 A100 GPUs
+num_devices,256,GPUs,stated,Figure 1: GPU utilization and memory usage metrics for a distributed model training on 256 A100 GPUs
+training_cost,,not_reported,not_reported,not_reported
+training_batch_size,,not_reported,not_reported,not_reported
+training_steps_or_epochs,,not_reported,not_reported,not_reported
+precision,,not_reported,not_reported,not_reported
+inference_latency,,not_reported,not_reported,not_reported
+inference_throughput,,not_reported,not_reported,not_reported
+peak_memory,,not_reported,not_reported,not_reported
+flops_or_macs,,not_reported,not_reported,not_reported
+num_inference_samples,,not_reported,not_reported,not_reported
+params,,not_reported,not_reported,not_reported
+context_lengths_evaluated,,not_reported,not_reported,not_reported
+horizon_lengths_evaluated,,not_reported,not_reported,not_reported
+inference_batch_size,,not_reported,not_reported,not_reported
diff --git a/result/per_paper/2502.00816/accuracy_efficiency.csv b/result/per_paper/2502.00816/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..c5960464a56a1c979a655cb8712bc3fa9e2771a4
--- /dev/null
+++ b/result/per_paper/2502.00816/accuracy_efficiency.csv
@@ -0,0 +1,42 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTm1,SundialSmall,MSE,0.317,
+accuracy,ETTm1,SundialBase,MSE,0.356,
+accuracy,ETTm1,SundialLarge,MSE,0.363,
+accuracy,ETTm2,SundialSmall,MSE,0.189,
+accuracy,ETTm2,SundialBase,MSE,0.277,
+accuracy,ETTm2,SundialLarge,MSE,0.205,
+accuracy,ETTh1,SundialSmall,MSE,0.369,
+accuracy,ETTh1,SundialBase,MSE,0.391,
+accuracy,ETTh1,SundialLarge,MSE,0.376,
+accuracy,ETTh2,SundialSmall,MSE,0.283,
+accuracy,ETTh2,SundialBase,MSE,0.342,
+accuracy,ETTh2,SundialLarge,MSE,0.294,
+accuracy,ECL,SundialSmall,MSE,0.141,
+accuracy,ECL,SundialBase,MSE,0.237,
+accuracy,ECL,SundialLarge,MSE,0.160,
+accuracy,Weather,SundialSmall,MSE,0.171,
+accuracy,Weather,SundialBase,MSE,0.225,
+accuracy,Weather,SundialLarge,MSE,0.220,
+accuracy,ETTh1,Sundial (94B),MSE,0.419,
+accuracy,ETTh1,Sundial (230B),MSE,0.411,
+accuracy,ETTh1,Sundial (1032B),MSE,0.434,
+accuracy,ETTh2,Sundial (94B),MSE,0.398,
+accuracy,ETTh2,Sundial (230B),MSE,0.333,
+accuracy,ETTh2,Sundial (1032B),MSE,0.387,
+accuracy,ETTm1,Sundial (94B),MSE,0.385,
+accuracy,ETTm1,Sundial (230B),MSE,0.336,
+accuracy,ETTm1,Sundial (1032B),MSE,0.377,
+accuracy,ETTm2,Sundial (94B),MSE,0.334,
+accuracy,ETTm2,Sundial (230B),MSE,0.258,
+accuracy,ETTm2,Sundial (1032B),MSE,0.320,
+accuracy,ECL,Sundial (94B),MSE,0.267,
+accuracy,ECL,Sundial (230B),MSE,0.169,
+accuracy,ECL,Sundial (1032B),MSE,0.265,
+accuracy,Weather,Sundial (94B),MSE,0.297,
+accuracy,Weather,Sundial (230B),MSE,0.234,
+accuracy,Weather,Sundial (1032B),MSE,0.270,
+accuracy,MASE,Sundial,MASE,0.673,
+accuracy,CRPS,Sundial,MASE,0.472,
+accuracy,Rank,Sundial,MASE,9.062,
+efficiency,Model Size,Sundial,Model Size,32,
+efficiency,Pre-training Scale,Sundial,Model Size,1032,
diff --git a/result/per_paper/2502.00816/accuracy_efficiency_traced.csv b/result/per_paper/2502.00816/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..dfdb6f6177d5531c46a0a6c960085da41ddc3be2
--- /dev/null
+++ b/result/per_paper/2502.00816/accuracy_efficiency_traced.csv
@@ -0,0 +1,42 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTm1,SundialSmall,MSE,0.317,,0,2,14
+accuracy,ETTm1,SundialBase,MSE,0.356,,0,2,15
+accuracy,ETTm1,SundialLarge,MSE,0.363,,0,2,16
+accuracy,ETTm2,SundialSmall,MSE,0.189,,0,7,14
+accuracy,ETTm2,SundialBase,MSE,0.277,,0,7,15
+accuracy,ETTm2,SundialLarge,MSE,0.205,,0,7,16
+accuracy,ETTh1,SundialSmall,MSE,0.369,,0,12,14
+accuracy,ETTh1,SundialBase,MSE,0.391,,0,12,15
+accuracy,ETTh1,SundialLarge,MSE,0.376,,0,12,16
+accuracy,ETTh2,SundialSmall,MSE,0.283,,0,17,14
+accuracy,ETTh2,SundialBase,MSE,0.342,,0,17,15
+accuracy,ETTh2,SundialLarge,MSE,0.294,,0,17,16
+accuracy,ECL,SundialSmall,MSE,0.141,,0,22,14
+accuracy,ECL,SundialBase,MSE,0.237,,0,22,15
+accuracy,ECL,SundialLarge,MSE,0.160,,0,22,16
+accuracy,Weather,SundialSmall,MSE,0.171,,0,27,14
+accuracy,Weather,SundialBase,MSE,0.225,,0,27,15
+accuracy,Weather,SundialLarge,MSE,0.220,,0,27,16
+accuracy,ETTh1,Sundial (94B),MSE,0.419,,2,2,8
+accuracy,ETTh1,Sundial (230B),MSE,0.411,,2,2,9
+accuracy,ETTh1,Sundial (1032B),MSE,0.434,,2,2,10
+accuracy,ETTh2,Sundial (94B),MSE,0.398,,2,3,8
+accuracy,ETTh2,Sundial (230B),MSE,0.333,,2,3,9
+accuracy,ETTh2,Sundial (1032B),MSE,0.387,,2,3,10
+accuracy,ETTm1,Sundial (94B),MSE,0.385,,2,4,8
+accuracy,ETTm1,Sundial (230B),MSE,0.336,,2,4,9
+accuracy,ETTm1,Sundial (1032B),MSE,0.377,,2,4,10
+accuracy,ETTm2,Sundial (94B),MSE,0.334,,2,5,8
+accuracy,ETTm2,Sundial (230B),MSE,0.258,,2,5,9
+accuracy,ETTm2,Sundial (1032B),MSE,0.320,,2,5,10
+accuracy,ECL,Sundial (94B),MSE,0.267,,2,6,8
+accuracy,ECL,Sundial (230B),MSE,0.169,,2,6,9
+accuracy,ECL,Sundial (1032B),MSE,0.265,,2,6,10
+accuracy,Weather,Sundial (94B),MSE,0.297,,2,7,8
+accuracy,Weather,Sundial (230B),MSE,0.234,,2,7,9
+accuracy,Weather,Sundial (1032B),MSE,0.270,,2,7,10
+accuracy,MASE,Sundial,MASE,0.673,,3,2,14
+accuracy,CRPS,Sundial,MASE,0.472,,3,3,14
+accuracy,Rank,Sundial,MASE,9.062,,3,4,14
+efficiency,Model Size,Sundial,Model Size,32,,7,2,1
+efficiency,Pre-training Scale,Sundial,Model Size,1032,,7,5,1
diff --git a/result/per_paper/2502.00816/components_architecture.csv b/result/per_paper/2502.00816/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5232fe7c050f45fd4bc2b3d5aa181ff9f0c16d2a
--- /dev/null
+++ b/result/per_paper/2502.00816/components_architecture.csv
@@ -0,0 +1,10 @@
+component,what_it_is,provenance,citation,evidence
+TimeFlow Loss,"A flow-matching-based loss function for predicting the next-patch distribution in continuous-valued time series, enabling native pre-training of Transformers without discrete tokenization.",proposed_here,"Liu et al., 2024","To predict the next-patch's distribution, we propose a TimeFlow Loss based on flow-matching, which facilitates native pre-training of Transformers on continuous-valued time series without discrete tokenization."
+Enhanced Transformer,"A modified Transformer architecture with minimal but crucial adaptations for time series foundation modeling, including RoPE, Pre-LN, FlashAttention, and KV Cache.",proposed_here,,"We enhance the Transformer with minimal but critical adaptations. We adopt RoPE, Pre-LN, FlashAttention, and KV Cache, which are crucial but generally neglected in the development of time series foundation models."
+Patch Tokenization,A method for tokenizing arbitrary-length input time series into patches for processing by the Transformer.,proposed_here,,We develop feasible patch tokenization for arbitrary-length input time series.
+RoPE (Rotary Position Embedding),"A position encoding technique for Transformers, adapted for time series modeling.",reused_cited,"Su et al., 2024","We adopt RoPE (Su et al., 2024), Pre-LN (Xiong et al., 2020), FlashAttention (Dao et al., 2022), and KV Cache (Pope et al., 2023)."
+Pre-LN (Pre-normalization),A normalization technique applied before layer operations in Transformer blocks.,reused_cited,"Xiong et al., 2020","We adopt RoPE (Su et al., 2024), Pre-LN (Xiong et al., 2020), FlashAttention (Dao et al., 2022), and KV Cache (Pope et al., 2023)."
+FlashAttention,An optimized attention mechanism for efficient computation in Transformers.,reused_cited,"Dao et al., 2022","We adopt RoPE (Su et al., 2024), Pre-LN (Xiong et al., 2020), FlashAttention (Dao et al., 2022), and KV Cache (Pope et al., 2023)."
+KV Cache (Key-Value Cache),A technique to store and reuse key-value pairs during inference for efficient long-context processing.,reused_cited,"Pope et al., 2023","We adopt RoPE (Su et al., 2024), Pre-LN (Xiong et al., 2020), FlashAttention (Dao et al., 2022), and KV Cache (Pope et al., 2023)."
+TimeBench Dataset,"A curated dataset with one trillion time points, comprising real-world and synthetic time series data for pre-training.",proposed_here,,"we curate TimeBench with one trillion time points, comprising mostly real-world datasets and synthetic data."
+Multi-patch Prediction,A pre-training strategy that reduces autoregression steps by predicting multiple patches simultaneously.,proposed_here,,We pre-train our models by multi-patch prediction to reduce autoregression steps.
diff --git a/result/per_paper/2502.00816/computational.csv b/result/per_paper/2502.00816/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..bc8a2ddc0119cd94f2268c03d709a6c7c3da464a
--- /dev/null
+++ b/result/per_paper/2502.00816/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA A100 GPU,,stated,All experiments are implemented using PyTorch [...] executed with 32 NVIDIA A100 GPUs.
+num_devices,32,,stated,All experiments are implemented using PyTorch [...] executed with 32 NVIDIA A100 GPUs.
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,1.0,ms,stated,Sundial (Base) | Pre-trained Models | ~1.0
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,20,,stated,sampling 20 predictions with each generated by 50 steps
+params,"32M, 128M, 444M",,stated,"Model table: SundialSmall (32M), SundialBase (128M), SundialLarge (444M)"
+context_lengths_evaluated,2880,tokens,stated,maximum context length of 2880; Model table: Context Length (T) = 2880
+horizon_lengths_evaluated,"{16, 720}",tokens,stated,"Model table: Prediction Length (F) = {16, 720}"
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2502.15637/accuracy_efficiency.csv b/result/per_paper/2502.15637/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..aa3edc57eb20adfa1e8883c015042afe6bb5bd00
--- /dev/null
+++ b/result/per_paper/2502.15637/accuracy_efficiency.csv
@@ -0,0 +1,466 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ACSF1,Mantis,accuracy,0.7433,0.0115
+accuracy,ACSF1,Mantis,accuracy,1.0,0.0
+accuracy,Adiac,Mantis,accuracy,0.7766,0.003
+accuracy,Adiac,Mantis,accuracy,0.5961,0.0051
+accuracy,AllGestureWiimoteX,Mantis,accuracy,0.7619,0.005
+accuracy,AllGestureWiimoteX,Mantis,accuracy,0.9624,0.0015
+accuracy,AllGestureWiimoteY,Mantis,accuracy,0.7948,0.0097
+accuracy,AllGestureWiimoteY,Mantis,accuracy,0.8516,0.0111
+accuracy,AllGestureWiimoteZ,Mantis,accuracy,0.73,0.01
+accuracy,AllGestureWiimoteZ,Mantis,accuracy,0.8087,0.0189
+accuracy,ArrowHead,Mantis,accuracy,0.821,0.0389
+accuracy,ArrowHead,Mantis,accuracy,0.7397,0.0362
+accuracy,BME,Mantis,accuracy,0.9956,0.0077
+accuracy,BME,Mantis,accuracy,0.9404,0.0072
+accuracy,Beef,Mantis,accuracy,0.7,0.0333
+accuracy,Beef,Mantis,accuracy,0.9333,0.0289
+accuracy,BeetleFly,Mantis,accuracy,0.8833,0.0764
+accuracy,BeetleFly,Mantis,accuracy,0.7662,0.0178
+accuracy,BirdChicken,Mantis,accuracy,0.9,0.05
+accuracy,BirdChicken,Mantis,accuracy,0.9552,0.001
+accuracy,CBF,Mantis,accuracy,0.9848,0.0074
+accuracy,CBF,Mantis,accuracy,0.5996,0.0209
+accuracy,Car,Mantis,accuracy,0.8722,0.0096
+accuracy,Car,Mantis,accuracy,0.8339,0.0139
+accuracy,Chinatown,Mantis,accuracy,0.9718,0.0017
+accuracy,Chinatown,Mantis,accuracy,0.4827,0.03
+accuracy,ChlorineConcentration,Mantis,accuracy,0.6971,0.0112
+accuracy,ChlorineConcentration,Mantis,accuracy,0.9531,0.0038
+accuracy,CinCECGTorso,Mantis,accuracy,0.7814,0.0173
+accuracy,CinCECGTorso,Mantis,accuracy,0.9068,0.0174
+accuracy,Coffee,Mantis,accuracy,1.0,0.0
+accuracy,Coffee,Mantis,accuracy,0.9086,0.0063
+accuracy,Computers,Mantis,accuracy,0.7813,0.0162
+accuracy,Computers,Mantis,accuracy,0.9216,0.0555
+accuracy,CricketX,Mantis,accuracy,0.7966,0.009
+accuracy,CricketX,Mantis,accuracy,0.9642,0.0024
+accuracy,CricketY,Mantis,accuracy,0.806,0.0141
+accuracy,CricketY,Mantis,accuracy,0.8889,0.0509
+accuracy,CricketZ,Mantis,accuracy,0.8043,0.0415
+accuracy,CricketZ,Mantis,accuracy,0.8162,0.0128
+accuracy,Crop,Mantis,accuracy,0.7444,0.0039
+accuracy,Crop,Mantis,accuracy,0.3214,0.0059
+accuracy,DiatomSizeReduction,Mantis,accuracy,0.9684,0.0136
+accuracy,DiatomSizeReduction,Mantis,accuracy,0.76,0.0529
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis,accuracy,0.7698,0.0259
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis,accuracy,1.0,0.0
+accuracy,DistalPhalanxOutlineCorrect,Mantis,accuracy,0.7717,0.0158
+accuracy,DistalPhalanxOutlineCorrect,Mantis,accuracy,0.9944,0.0096
+accuracy,DistalPhalanxTW,Mantis,accuracy,0.6954,0.0231
+accuracy,DistalPhalanxTW,Mantis,accuracy,0.8455,0.0246
+accuracy,DodgerLoopDay,Mantis,accuracy,0.625,0.025
+accuracy,DodgerLoopDay,Mantis,accuracy,0.8763,0.0209
+accuracy,DodgerLoopGame,Mantis,accuracy,0.8841,0.0145
+accuracy,DodgerLoopGame,Mantis,accuracy,0.7772,0.0314
+accuracy,DodgerLoopWeekend,Mantis,accuracy,0.9783,0.0
+accuracy,DodgerLoopWeekend,Mantis,accuracy,0.4916,0.0111
+accuracy,ECG200,Mantis,accuracy,0.8567,0.0058
+accuracy,ECG200,Mantis,accuracy,0.7133,0.0231
+accuracy,ECG5000,Mantis,accuracy,0.9335,0.0081
+accuracy,ECG5000,Mantis,accuracy,0.4782,0.0041
+accuracy,ECGFiveDays,Mantis,accuracy,0.9148,0.0292
+accuracy,ECGFiveDays,Mantis,accuracy,0.92,0.0
+accuracy,Earthquakes,Mantis,accuracy,0.7482,0.
+accuracy,Earthquakes,Mantis,accuracy,1.0,0.0
+accuracy,ElectricDevices,Mantis,accuracy,0.7454,0.0049
+accuracy,ElectricDevices,Mantis,accuracy,0.8906,0.0059
+accuracy,FaceAll,Mantis,accuracy,0.8308,0.0031
+accuracy,FaceAll,Mantis,accuracy,0.7929,0.0031
+accuracy,FaceFour,Mantis,accuracy,0.9773,0.0114
+accuracy,FaceFour,Mantis,accuracy,0.9978,0.0038
+accuracy,FacesUCR,Mantis,accuracy,0.9148,0.0071
+accuracy,FacesUCR,Mantis,accuracy,0.8142,0.0113
+accuracy,FiftyWords,Mantis,accuracy,0.8139,0.0083
+accuracy,FiftyWords,Mantis,accuracy,0.8996,0.0112
+accuracy,Fish,Mantis,accuracy,0.9714,0.0099
+accuracy,Fish,Mantis,accuracy,0.9667,0.0078
+accuracy,FordA,Mantis,accuracy,0.9343,0.0046
+accuracy,FordA,Mantis,accuracy,0.9712,0.0042
+accuracy,FordB,Mantis,accuracy,0.7979,0.0135
+accuracy,FordB,Mantis,accuracy,0.9889,0.0017
+accuracy,FreezerRegularTrain,Mantis,accuracy,0.9927,0.0041
+accuracy,FreezerRegularTrain,Mantis,accuracy,0.9944,0.0038
+accuracy,FreezerSmallTrain,Mantis,accuracy,0.9667,0.0098
+accuracy,FreezerSmallTrain,Mantis,accuracy,0.9664,0.0067
+accuracy,Fungi,Mantis,accuracy,0.7778,0.0031
+accuracy,ACSF1,Mantis-RF,accuracy,0.6133,0.0208
+accuracy,ACSF1,Mantis-Head,accuracy,0.62,0.0265
+accuracy,ACSF1,Mantis-Scratch,accuracy,0.6433,0.0451
+accuracy,ACSF1,Mantis-Full,accuracy,0.7467,0.0416
+accuracy,Adiac,Mantis-RF,accuracy,0.7332,0.0039
+accuracy,Adiac,Mantis-Head,accuracy,0.5857,0.0051
+accuracy,Adiac,Mantis-Scratch,accuracy,0.7613,0.0129
+accuracy,Adiac,Mantis-Full,accuracy,0.8295,0.0082
+accuracy,AllGestureWiimoteX,Mantis-RF,accuracy,0.6705,0.0044
+accuracy,AllGestureWiimoteX,Mantis-Head,accuracy,0.551,0.0008
+accuracy,AllGestureWiimoteX,Mantis-Scratch,accuracy,0.6962,0.0058
+accuracy,AllGestureWiimoteX,Mantis-Full,accuracy,0.8029,0.0065
+accuracy,AllGestureWiimoteY,Mantis-RF,accuracy,0.6671,0.0057
+accuracy,AllGestureWiimoteY,Mantis-Head,accuracy,0.5614,0.0052
+accuracy,AllGestureWiimoteY,Mantis-Scratch,accuracy,0.739,0.0008
+accuracy,AllGestureWiimoteY,Mantis-Full,accuracy,0.8129,0.0089
+accuracy,AllGestureWiimoteZ,Mantis-RF,accuracy,0.6695,0.0033
+accuracy,AllGestureWiimoteZ,Mantis-Head,accuracy,0.5467,0.0008
+accuracy,AllGestureWiimoteZ,Mantis-Scratch,accuracy,0.6367,0.0246
+accuracy,AllGestureWiimoteZ,Mantis-Full,accuracy,0.7481,0.0095
+accuracy,ArrowHead,Mantis-RF,accuracy,0.7105,0.0175
+accuracy,ArrowHead,Mantis-Head,accuracy,0.7276,0.0119
+accuracy,ArrowHead,Mantis-Scratch,accuracy,0.8,0.0286
+accuracy,ArrowHead,Mantis-Full,accuracy,0.8229,0.0297
+accuracy,BME,Mantis-RF,accuracy,0.9311,0.0038
+accuracy,BME,Mantis-Head,accuracy,0.8756,0.0139
+accuracy,BME,Mantis-Scratch,accuracy,0.9933,0.0
+accuracy,BME,Mantis-Full,accuracy,1.0,0.0
+accuracy,Beef,Mantis-RF,accuracy,0.6556,0.0509
+accuracy,Beef,Mantis-Head,accuracy,0.6222,0.0385
+accuracy,Beef,Mantis-Scratch,accuracy,0.8222,0.0694
+accuracy,Beef,Mantis-Full,accuracy,0.8667,0.0
+accuracy,BeetleFly,Mantis-RF,accuracy,0.85,0.05
+accuracy,BeetleFly,Mantis-Head,accuracy,0.9,0.0
+accuracy,BeetleFly,Mantis-Scratch,accuracy,0.7,0.0
+accuracy,BeetleFly,Mantis-Full,accuracy,0.9,0.05
+accuracy,BirdChicken,Mantis-RF,accuracy,1.0,0.0
+accuracy,BirdChicken,Mantis-Head,accuracy,1.0,0.0
+accuracy,BirdChicken,Mantis-Scratch,accuracy,0.9,0.0
+accuracy,BirdChicken,Mantis-Full,accuracy,1.0,0.0
+accuracy,CBF,Mantis-RF,accuracy,0.993,0.0013
+accuracy,CBF,Mantis-Head,accuracy,0.9941,0.0006
+accuracy,CBF,Mantis-Scratch,accuracy,0.9922,0.008
+accuracy,CBF,Mantis-Full,accuracy,0.9996,0.0006
+accuracy,Car,Mantis-RF,accuracy,0.7722,0.0419
+accuracy,Car,Mantis-Head,accuracy,0.7778,0.0192
+accuracy,Car,Mantis-Scratch,accuracy,0.8,0.0
+accuracy,Car,Mantis-Full,accuracy,0.9,0.0289
+accuracy,Chinatown,Mantis-RF,accuracy,0.8737,0.0168
+accuracy,Chinatown,Mantis-Head,accuracy,0.8785,0.0094
+accuracy,Chinatown,Mantis-Scratch,accuracy,0.964,0.0017
+accuracy,Chinatown,Mantis-Full,accuracy,0.9776,0.0067
+accuracy,ChlorineConcentration,Mantis-RF,accuracy,0.6806,0.0019
+accuracy,ChlorineConcentration,Mantis-Head,accuracy,0.5704,0.0011
+accuracy,ChlorineConcentration,Mantis-Scratch,accuracy,0.753,0.0056
+accuracy,ChlorineConcentration,Mantis-Full,accuracy,0.7788,0.0072
+accuracy,CinCECGTorso,Mantis-RF,accuracy,0.6611,0.0036
+accuracy,CinCECGTorso,Mantis-Head,accuracy,0.7179,0.0048
+accuracy,CinCECGTorso,Mantis-Scratch,accuracy,0.7901,0.0198
+accuracy,CinCECGTorso,Mantis-Full,accuracy,0.856,0.0077
+accuracy,Coffee,Mantis-RF,accuracy,0.9524,0.0206
+accuracy,Coffee,Mantis-Head,accuracy,0.9643,0.0
+accuracy,Coffee,Mantis-Scratch,accuracy,1.0,0.0
+accuracy,Coffee,Mantis-Full,accuracy,1.0,0.0
+accuracy,Computers,Mantis-RF,accuracy,0.7373,0.0092
+accuracy,Computers,Mantis-Head,accuracy,0.728,0.0069
+accuracy,Computers,Mantis-Scratch,accuracy,0.7533,0.0162
+accuracy,Computers,Mantis-Full,accuracy,0.7773,0.0046
+accuracy,CricketX,Mantis-RF,accuracy,0.7368,0.0171
+accuracy,CricketX,Mantis-Head,accuracy,0.6966,0.003
+accuracy,CricketX,Mantis-Scratch,accuracy,0.7812,0.0146
+accuracy,CricketX,Mantis-Full,accuracy,0.8282,0.0133
+accuracy,CricketY,Mantis-RF,accuracy,0.7504,0.0065
+accuracy,CricketY,Mantis-Head,accuracy,0.7051,0.0026
+accuracy,CricketY,Mantis-Scratch,accuracy,0.8,0.0026
+accuracy,CricketY,Mantis-Full,accuracy,0.8385,0.0051
+accuracy,CricketZ,Mantis-RF,accuracy,0.7900,0.0039
+accuracy,CricketZ,Mantis-Head,accuracy,0.7333,0.0077
+accuracy,CricketZ,Mantis-Scratch,accuracy,0.7991,0.0097
+accuracy,CricketZ,Mantis-Full,accuracy,0.8521,0.0082
+accuracy,Crop,Mantis-RF,accuracy,0.6756,0.0018
+accuracy,Crop,Mantis-Head,accuracy,0.6745,0.0001
+accuracy,Crop,Mantis-Scratch,accuracy,0.7566,0.0063
+accuracy,Crop,Mantis-Full,accuracy,0.7702,0.0005
+accuracy,DiatomSizeReduction,Mantis-RF,accuracy,0.8845,0.0019
+accuracy,DiatomSizeReduction,Mantis-Head,accuracy,0.8911,0.0191
+accuracy,DiatomSizeReduction,Mantis-Scratch,accuracy,0.9031,0.0147
+accuracy,DiatomSizeReduction,Mantis-Full,accuracy,0.9346,0.0033
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-RF,accuracy,0.789,0.015
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-Head,accuracy,0.8106,0.0042
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-Scratch,accuracy,0.7506,0.011
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-Full,accuracy,0.7938,0.0042
+accuracy,DistalPhalanxOutlineCorrect,Mantis-RF,accuracy,0.75,0.0126
+accuracy,DistalPhalanxOutlineCorrect,Mantis-Head,accuracy,0.7729,0.0055
+accuracy,DistalPhalanxOutlineCorrect,Mantis-Scratch,accuracy,0.8056,0.0091
+accuracy,DistalPhalanxOutlineCorrect,Mantis-Full,accuracy,0.7935,0.0036
+accuracy,DistalPhalanxTW,Mantis-RF,accuracy,0.6859,0.011
+accuracy,DistalPhalanxTW,Mantis-Head,accuracy,0.6763,0.0125
+accuracy,DistalPhalanxTW,Mantis-Scratch,accuracy,0.7266,0.0125
+accuracy,DistalPhalanxTW,Mantis-Full,accuracy,0.7338,0.019
+accuracy,DodgerLoopDay,Mantis-RF,accuracy,0.55,0.0217
+accuracy,DodgerLoopDay,Mantis-Head,accuracy,0.5292,0.026
+accuracy,DodgerLoopDay,Mantis-Scratch,accuracy,0.6208,0.0191
+accuracy,DodgerLoopDay,Mantis-Full,accuracy,0.6375,0.0125
+accuracy,DodgerLoopGame,Mantis-RF,accuracy,0.7585,0.0221
+accuracy,DodgerLoopGame,Mantis-Head,accuracy,0.7222,0.0111
+accuracy,DodgerLoopGame,Mantis-Scratch,accuracy,0.843,0.0084
+accuracy,DodgerLoopGame,Mantis-Full,accuracy,0.9541,0.0042
+accuracy,DodgerLoopWeekend,Mantis-RF,accuracy,0.9517,0.0084
+accuracy,DodgerLoopWeekend,Mantis-Head,accuracy,0.9783,0.0126
+accuracy,DodgerLoopWeekend,Mantis-Scratch,accuracy,0.9855,0.0
+accuracy,DodgerLoopWeekend,Mantis-Full,accuracy,0.9855,0.0
+accuracy,ECG200,Mantis-RF,accuracy,0.82,0.01
+accuracy,ECG200,Mantis-Head,accuracy,0.8033,0.0252
+accuracy,ECG200,Mantis-Scratch,accuracy,0.8467,0.0115
+accuracy,ECG200,Mantis-Full,accuracy,0.9133,0.0115
+accuracy,ECG5000,Mantis-RF,accuracy,0.9211,0.001
+accuracy,ECG5000,Mantis-Head,accuracy,0.9185,0.0038
+accuracy,ECG5000,Mantis-Scratch,accuracy,0.9411,0.0008
+accuracy,ECG5000,Mantis-Full,accuracy,0.9459,0.0014
+accuracy,ECGFiveDays,Mantis-RF,accuracy,0.909,0.0218
+accuracy,ECGFiveDays,Mantis-Head,accuracy,0.8668,0.0401
+accuracy,ECGFiveDays,Mantis-Scratch,accuracy,0.7855,0.0169
+accuracy,ECGFiveDays,Mantis-Full,accuracy,0.9694,0.014
+accuracy,EOGHorizontalSignal,Mantis-RF,accuracy,0.5875,0.0089
+accuracy,EOGHorizontalSignal,Mantis-Head,accuracy,0.5783,0.0064
+accuracy,EOGHorizontalSignal,Mantis-Scratch,accuracy,0.6703,0.0194
+accuracy,EOGHorizontalSignal,Mantis-Full,accuracy,0.6934,0.012
+accuracy,EOGVerticalSignal,Mantis-RF,accuracy,0.4751,0.0
+accuracy,EOGVerticalSignal,Mantis-Head,accuracy,0.4825,0.0105
+accuracy,EOGVerticalSignal,Mantis-Scratch,accuracy,0.5396,0.0058
+accuracy,EOGVerticalSignal,Mantis-Full,accuracy,0.5773,0.0073
+accuracy,Earthquakes,Mantis-RF,accuracy,0.7482,0.0
+accuracy,Earthquakes,Mantis-Head,accuracy,0.7506,0.0042
+accuracy,Earthquakes,Mantis-Scratch,accuracy,0.7866,0.015
+accuracy,Earthquakes,Mantis-Full,accuracy,0.7602,0.015
+accuracy,ElectricDevices,Mantis-RF,accuracy,0.7226,0.0026
+accuracy,ElectricDevices,Mantis-Head,accuracy,0.7064,0.001
+accuracy,ElectricDevices,Mantis-Scratch,accuracy,0.7448,0.0044
+accuracy,ElectricDevices,Mantis-Full,accuracy,0.7643,0.005
+accuracy,EthanolLevel,Mantis-RF,accuracy,0.2993,0.011
+accuracy,EthanolLevel,Mantis-Head,accuracy,0.324,0.006
+accuracy,EthanolLevel,Mantis-Scratch,accuracy,0.6987,0.0194
+accuracy,EthanolLevel,Mantis-Full,accuracy,0.7393,0.1052
+accuracy,FaceAll,Mantis-RF,accuracy,0.7815,0.0074
+accuracy,FaceAll,Mantis-Head,accuracy,0.8053,0.0062
+accuracy,FaceAll,Mantis-Scratch,accuracy,0.9542,0.0036
+accuracy,FaceAll,Mantis-Full,accuracy,0.8554,0.0343
+accuracy,FaceFour,Mantis-RF,accuracy,0.9508,0.0066
+accuracy,FaceFour,Mantis-Head,accuracy,0.9508,0.0174
+accuracy,FaceFour,Mantis-Scratch,accuracy,0.8485,0.0347
+accuracy,FaceFour,Mantis-Full,accuracy,0.9811,0.0066
+accuracy,FacesUCR,Mantis-RF,accuracy,0.8354,0.0054
+accuracy,FacesUCR,Mantis-Head,accuracy,0.8437,0.0104
+accuracy,FacesUCR,Mantis-Scratch,accuracy,0.8816,0.0054
+accuracy,FacesUCR,Mantis-Full,accuracy,0.942,0.0061
+accuracy,FiftyWords,Mantis-RF,accuracy,0.6462,0.0096
+accuracy,FiftyWords,Mantis-Head,accuracy,0.7026,0.0178
+accuracy,FiftyWords,Mantis-Scratch,accuracy,0.8176,0.0038
+accuracy,FiftyWords,Mantis-Full,accuracy,0.863,0.0013
+accuracy,Fish,Mantis-RF,accuracy,0.9333,0.0066
+accuracy,Fish,Mantis-Head,accuracy,0.9333,0.0033
+accuracy,Fish,Mantis-Scratch,accuracy,0.939,0.0087
+accuracy,Fish,Mantis-Full,accuracy,0.979,0.0087
+accuracy,FordA,Mantis-RF,accuracy,0.8581,0.0048
+accuracy,FordA,Mantis-Head,accuracy,0.8881,0.0004
+accuracy,FordA,Mantis-Scratch,accuracy,0.9301,0.0031
+accuracy,FordA,Mantis-Full,accuracy,0.9477,0.0013
+accuracy,FordB,Mantis-RF,accuracy,0.7305,0.0031
+accuracy,FordB,Mantis-Head,accuracy,0.7704,0.0033
+accuracy,FordB,Mantis-Scratch,accuracy,0.8115,0.0036
+accuracy,FordB,Mantis-Full,accuracy,0.8477,0.004
+accuracy,FreezerRegularTrain,Mantis-RF,accuracy,0.9374,0.0043
+accuracy,FreezerRegularTrain,Mantis-Head,accuracy,0.8767,0.0088
+accuracy,FreezerRegularTrain,Mantis-Scratch,accuracy,0.9971,0.0004
+accuracy,FreezerRegularTrain,Mantis-Full,accuracy,0.9967,0.0014
+accuracy,FreezerSmallTrain,Mantis-RF,accuracy,0.7942,0.0059
+accuracy,FreezerSmallTrain,Mantis-Head,accuracy,0.8092,0.0041
+accuracy,FreezerSmallTrain,Mantis-Scratch,accuracy,0.8601,0.073
+accuracy,FreezerSmallTrain,Mantis-Full,accuracy,0.9648,0.0219
+accuracy,Fungi,Mantis-RF,accuracy,0.8262,0.0164
+accuracy,Fungi,Mantis-Head,accuracy,0.8477,0.0224
+accuracy,Fungi,Mantis-Scratch,accuracy,0.948,0.0164
+accuracy,Fungi,Mantis-Full,accuracy,0.9892,0.0108
+accuracy,GestureMidAirD1,Mantis-RF,accuracy,0.659,0.0044
+accuracy,GestureMidAirD1,Mantis-Head,accuracy,0.6026,0.0044
+accuracy,GestureMidAirD1,Mantis-Scratch,accuracy,0.7103,0.0089
+accuracy,GestureMidAirD1,Mantis-Full,accuracy,0.7949,0.0044
+accuracy,GestureMidAirD2,Mantis-RF,accuracy,0.6154,0.0077
+accuracy,GestureMidAirD2,Mantis-Head,accuracy,0.5282,0.0118
+accuracy,GestureMidAirD2,Mantis-Scratch,accuracy,0.6974,0.0222
+accuracy,GestureMidAirD2,Mantis-Full,accuracy,0.7205,0.0247
+accuracy,GestureMidAirD3,Mantis-RF,accuracy,0.3282,0.016
+accuracy,GestureMidAirD3,Mantis-Head,accuracy,0.3154,0.0266
+accuracy,GestureMidAirD3,Mantis-Scratch,accuracy,0.4359,0.016
+accuracy,GestureMidAirD3,Mantis-Full,accuracy,0.4795,0.0178
+accuracy,GesturePebbleZ1,Mantis-RF,accuracy,0.9283,0.0034
+accuracy,GesturePebbleZ1,Mantis-Head,accuracy,0.9186,0.0116
+accuracy,GesturePebbleZ1,Mantis-Scratch,accuracy,0.9322,0.0089
+accuracy,GesturePebbleZ1,Mantis-Full,accuracy,0.9574,0.0146
+accuracy,GesturePebbleZ2,Mantis-RF,accuracy,0.9219,0.0256
+accuracy,GesturePebbleZ2,Mantis-Head,accuracy,0.9515,0.0097
+accuracy,GesturePebbleZ2,Mantis-Scratch,accuracy,0.8882,0.0193
+accuracy,GesturePebbleZ2,Mantis-Full,accuracy,0.9451,0.0159
+accuracy,GunPoint,Mantis-RF,accuracy,0.98,0.0067
+accuracy,GunPoint,Mantis-Head,accuracy,0.9822,0.0038
+accuracy,GunPoint,Mantis-Scratch,accuracy,0.9867,0.0133
+accuracy,GunPoint,Mantis-Full,accuracy,0.9978,0.0038
+accuracy,GunPointAgeSpan,Mantis-RF,accuracy,0.9905,0.0
+accuracy,GunPointAgeSpan,Mantis-Head,accuracy,0.9821,0.0037
+accuracy,GunPointAgeSpan,Mantis-Scratch,accuracy,0.9947,0.0048
+accuracy,GunPointAgeSpan,Mantis-Full,accuracy,0.9989,0.0018
+accuracy,GunPointMaleVersusFemale,Mantis-RF,accuracy,0.9958,0.0018
+accuracy,GunPointMaleVersusFemale,Mantis-Head,accuracy,0.9863,0.0048
+accuracy,GunPointMaleVersusFemale,Mantis-Scratch,accuracy,0.9947,0.0018
+accuracy,GunPointMaleVersusFemale,Mantis-Full,accuracy,1.0,0.0
+accuracy,GunPointOldVersusYoung,Mantis-RF,accuracy,0.9968,0.0
+accuracy,GunPointOldVersusYoung,Mantis-Head,accuracy,0.9968,0.0
+accuracy,GunPointOldVersusYoung,Mantis-Scratch,accuracy,0.9968,0.0
+accuracy,GunPointOldVersusYoung,Mantis-Full,accuracy,1.0,0.0
+accuracy,Ham,Mantis-RF,accuracy,0.673,0.0145
+accuracy,Ham,Mantis-Head,accuracy,0.7143,0.0252
+accuracy,Ham,Mantis-Scratch,accuracy,0.8159,0.0198
+accuracy,Ham,Mantis-Full,accuracy,0.7016,0.0198
+accuracy,HandOutlines,Mantis-RF,accuracy,0.9162,0.0072
+accuracy,HandOutlines,Mantis-Head,accuracy,0.8928,0.0068
+accuracy,HandOutlines,Mantis-Scratch,accuracy,0.945,0.0056
+accuracy,HandOutlines,Mantis-Full,accuracy,0.955,0.0016
+accuracy,Haptics,Mantis-RF,accuracy,0.4968,0.0032
+accuracy,Haptics,Mantis-Head,accuracy,0.4697,0.0131
+accuracy,Haptics,Mantis-Scratch,accuracy,0.5032,0.0065
+accuracy,Haptics,Mantis-Full,accuracy,0.5747,0.0234
+accuracy,Herring,Mantis-RF,accuracy,0.6667,0.0239
+accuracy,Herring,Mantis-Head,accuracy,0.7188,0.0271
+accuracy,Herring,Mantis-Scratch,accuracy,0.6458,0.0239
+accuracy,Herring,Mantis-Full,accuracy,0.724,0.009
+accuracy,HouseTwenty,Mantis-RF,accuracy,0.9412,0.0
+accuracy,HouseTwenty,Mantis-Head,accuracy,0.9496,0.0084
+accuracy,HouseTwenty,Mantis-Scratch,accuracy,0.9356,0.0049
+accuracy,HouseTwenty,Mantis-Full,accuracy,0.9776,0.0097
+accuracy,InlineSkate,Mantis-RF,accuracy,0.363,0.0136
+accuracy,InlineSkate,Mantis-Head,accuracy,0.3182,0.0101
+accuracy,InlineSkate,Mantis-Scratch,accuracy,0.403,0.0211
+accuracy,InlineSkate,Mantis-Full,accuracy,0.4515,0.0132
+accuracy,InsectEPGRegularTrain,Mantis-RF,accuracy,1.0,0.0
+accuracy,InsectEPGRegularTrain,Mantis-Head,accuracy,1.0,0.0
+accuracy,InsectEPGRegularTrain,Mantis-Scratch,accuracy,0.992,0.0
+accuracy,InsectEPGRegularTrain,Mantis-Full,accuracy,1.0,0.0
+accuracy,InsectEPGSmallTrain,Mantis-RF,accuracy,1.0,0.0
+accuracy,InsectEPGSmallTrain,Mantis-Head,accuracy,1.0,0.0
+accuracy,InsectEPGSmallTrain,Mantis-Scratch,accuracy,0.9665,0.0061
+accuracy,InsectEPGSmallTrain,Mantis-Full,accuracy,1.0,0.0
+accuracy,InsectWingbeatSound,Mantis-RF,accuracy,0.519,0.0044
+accuracy,InsectWingbeatSound,Mantis-Head,accuracy,0.4791,0.0099
+accuracy,InsectWingbeatSound,Mantis-Scratch,accuracy,0.6222,0.0066
+accuracy,InsectWingbeatSound,Mantis-Full,accuracy,0.6246,0.0068
+accuracy,Blink,Mantis-RF,accuracy,1.0,0.0
+accuracy,Blink,Mantis-Head,accuracy,0.9963,0.0046
+accuracy,Blink,Mantis-Scratch,accuracy,0.9896,0.0056
+accuracy,Blink,Mantis-Full,accuracy,1.0,0.0
+accuracy,MotionSenseHAR,Mantis-RF,accuracy,1.0,0.0
+accuracy,MotionSenseHAR,Mantis-Head,accuracy,0.9962,0.0
+accuracy,MotionSenseHAR,Mantis-Scratch,accuracy,0.9887,0.0
+accuracy,MotionSenseHAR,Mantis-Full,accuracy,1.0,0.0
+accuracy,SharePriceIncrease,Mantis-RF,accuracy,0.6874,0.0
+accuracy,SharePriceIncrease,Mantis-Head,accuracy,0.6746,0.0012
+accuracy,SharePriceIncrease,Mantis-Scratch,accuracy,0.6891,0.0039
+accuracy,SharePriceIncrease,Mantis-Full,accuracy,0.6874,0.0027
+accuracy,ArticularyWordRecognition,Mantis-RF,accuracy,0.9933,0.0033
+accuracy,ArticularyWordRecognition,Mantis-Head,accuracy,0.9933,0.0
+accuracy,ArticularyWordRecognition,Mantis-Scratch,accuracy,0.9967,0.0
+accuracy,ArticularyWordRecognition,Mantis-Full,accuracy,0.9933,0.0
+accuracy,BasicMotions,Mantis-RF,accuracy,1.0,0.0
+accuracy,BasicMotions,Mantis-Head,accuracy,1.0,0.0
+accuracy,BasicMotions,Mantis-Scratch,accuracy,1.0,0.0
+accuracy,BasicMotions,Mantis-Full,accuracy,1.0,0.0
+accuracy,CharacterTrajectories,Mantis-RF,accuracy,0.9396,0.0008
+accuracy,CharacterTrajectories,Mantis-Head,accuracy,0.9673,0.0021
+accuracy,CharacterTrajectories,Mantis-Scratch,accuracy,0.9838,0.0029
+accuracy,CharacterTrajectories,Mantis-Full,accuracy,0.9928,0.0004
+accuracy,Cricket,Mantis-RF,accuracy,1.0,0.0
+accuracy,Cricket,Mantis-Head,accuracy,0.9907,0.008
+accuracy,Cricket,Mantis-Scratch,accuracy,0.9861,0.0
+accuracy,Cricket,Mantis-Full,accuracy,1.0,0.0
+accuracy,ERing,Mantis-RF,accuracy,0.9407,0.0098
+accuracy,ERing,Mantis-Head,accuracy,0.9296,0.0192
+accuracy,ERing,Mantis-Scratch,accuracy,0.9432,0.0077
+accuracy,ERing,Mantis-Full,accuracy,0.9926,0.0074
+accuracy,EigenWorms,Mantis-RF,accuracy,0.7532,0.0159
+accuracy,EigenWorms,Mantis-Head,accuracy,0.7379,0.0233
+accuracy,EigenWorms,Mantis-Scratch,accuracy,0.7023,0.0202
+accuracy,EigenWorms,Mantis-Full,accuracy,0.8372,0.0044
+accuracy,Epilepsy,Mantis-RF,accuracy,0.9952,0.0042
+accuracy,Epilepsy,Mantis-Head,accuracy,0.9928,0.0
+accuracy,Epilepsy,Mantis-Scratch,accuracy,1.0,0.0
+accuracy,Epilepsy,Mantis-Full,accuracy,1.0,0.0
+accuracy,EthanolConcentration,Mantis-RF,accuracy,0.27,0.0101
+accuracy,EthanolConcentration,Mantis-Head,accuracy,0.3156,0.0076
+accuracy,EthanolConcentration,Mantis-Scratch,accuracy,0.3321,0.0154
+accuracy,EthanolConcentration,Mantis-Full,accuracy,0.4208,0.0195
+accuracy,HandMovementDirection,Mantis-RF,accuracy,0.2117,0.0413
+accuracy,HandMovementDirection,Mantis-Head,accuracy,0.3829,0.0639
+accuracy,HandMovementDirection,Mantis-Scratch,accuracy,0.5135,0.0234
+accuracy,HandMovementDirection,Mantis-Full,accuracy,0.4009,0.0206
+accuracy,ACSF1,Mantis,performance,0.6133,
+accuracy,Adiac,Mantis,performance,0.7332,
+accuracy,AllGestureWiimoteX,Mantis,performance,0.6705,
+accuracy,AllGestureWiimoteY,Mantis,performance,0.6671,
+accuracy,AllGestureWiimoteZ,Mantis,performance,0.6695,
+accuracy,ArrowHead,Mantis,performance,0.7105,
+accuracy,BME,Mantis,performance,0.9311,
+accuracy,Beef,Mantis,performance,0.6556,
+accuracy,BeetleFly,Mantis,performance,0.85,
+accuracy,BirdChicken,Mantis,performance,1.0,
+accuracy,CBF,Mantis,performance,0.993,
+accuracy,Car,Mantis,performance,0.7722,
+accuracy,Chinatown,Mantis,performance,0.8737,
+accuracy,ChlorineConcentration,Mantis,performance,0.6806,
+accuracy,CinCECGTorso,Mantis,performance,0.6611,
+accuracy,Coffee,Mantis,performance,0.9524,
+accuracy,Computers,Mantis,performance,0.7373,
+accuracy,CricketX,Mantis,performance,0.7368,
+accuracy,CricketY,Mantis,performance,0.7504,
+accuracy,CricketZ,Mantis,performance,0.7906,
+accuracy,Crop,Mantis,performance,0.6756,
+accuracy,DiatomSizeReduction,Mantis,performance,0.8845,
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis,performance,0.789,
+accuracy,DistalPhalanxOutlineCorrect,Mantis,performance,0.75,
+accuracy,DistalPhalanxTW,Mantis,performance,0.6859,
+accuracy,DodgerLoopDay,Mantis,performance,0.55,
+accuracy,DodgerLoopGame,Mantis,performance,0.7585,
+accuracy,DodgerLoopWeekend,Mantis,performance,0.9517,
+accuracy,ECG200,Mantis,performance,0.82,
+accuracy,ECG5000,Mantis,performance,0.9211,
+accuracy,ECGFiveDays,Mantis,performance,0.909,
+accuracy,EOGHorizontalSignal,Mantis,performance,0.5875,
+accuracy,EOGVerticalSignal,Mantis,performance,0.4751,
+accuracy,Earthquakes,Mantis,performance,0.7482,
+accuracy,ElectricDevices,Mantis,performance,0.7226,
+accuracy,EthanolLevel,Mantis,performance,0.2993,
+accuracy,FaceAll,Mantis,performance,0.7815,
+accuracy,FaceFour,Mantis,performance,0.9508,
+accuracy,FacesUCR,Mantis,performance,0.8354,
+accuracy,FiftyWords,Mantis,performance,0.6462,
+accuracy,Fish,Mantis,performance,0.9333,
+accuracy,FordA,Mantis,performance,0.8581,
+accuracy,FordB,Mantis,performance,0.7305,
+accuracy,FreezerRegularTrain,Mantis,performance,0.9374,
+accuracy,FreezerSmallTrain,Mantis,performance,0.7942,
+accuracy,Fungi,Mantis,performance,0.8262,
+accuracy,GestureMidAirD2,Mantis,performance,0.6154,
+accuracy,GestureMidAirD3,Mantis,performance,0.3282,
+accuracy,GesturePebbleZ1,Mantis,performance,0.9283,
+accuracy,GesturePebbleZ2,Mantis,performance,0.9219,
+accuracy,GunPoint,Mantis,performance,0.98,
+accuracy,GunPointAgeSpan,Mantis,performance,0.9905,
+accuracy,GunPointMaleVersusFemale,Mantis,performance,0.9958,
+accuracy,GunPointOldVersusYoung,Mantis,performance,0.9968,
+accuracy,Ham,Mantis,performance,0.673,
+accuracy,HandOutlines,Mantis,performance,0.9162,
+accuracy,Haptics,Mantis,performance,0.4968,
+accuracy,Herring,Mantis,performance,0.6667,
+accuracy,HouseTwenty,Mantis,performance,0.9412,
+accuracy,InlineSkate,Mantis,performance,0.363,
+accuracy,InsectWingbeatSound,Mantis,performance,0.519,
+accuracy,Blink,Mantis,performance,0.9993,
+accuracy,EMOPain,Mantis,performance,0.8385,
+accuracy,MotionSenseHAR,Mantis,performance,0.9975,
+accuracy,SharePriceIncrease,Mantis,performance,0.6836,
+accuracy,ArticularyWordRecognition,Mantis,performance,0.993,
+accuracy,CharacterTrajectories,Mantis,performance,0.94,
+accuracy,Cricket,Mantis,performance,1.0,
+accuracy,DuckDuckGeese,Mantis,performance,0.407,
+accuracy,ERing,Mantis,performance,0.941,
+accuracy,EigenWorms,Mantis,performance,0.753,
+accuracy,Epilepsy,Mantis,performance,0.995,
+accuracy,EthanolConcentration,Mantis,performance,0.27,
+accuracy,FaceDetection,Mantis,performance,0.526,
+accuracy,FingerMovements,Mantis,performance,0.54,
+accuracy,HandMovementDirection,Mantis,performance,0.212,
diff --git a/result/per_paper/2502.15637/accuracy_efficiency_traced.csv b/result/per_paper/2502.15637/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..06bc9fb7692e00b621f565a7825b755e2b8db00a
--- /dev/null
+++ b/result/per_paper/2502.15637/accuracy_efficiency_traced.csv
@@ -0,0 +1,466 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ACSF1,Mantis,accuracy,0.7433,0.0115,0,1,5
+accuracy,ACSF1,Mantis,accuracy,1.0,0.0,0,1,11
+accuracy,Adiac,Mantis,accuracy,0.7766,0.003,0,2,5
+accuracy,Adiac,Mantis,accuracy,0.5961,0.0051,0,2,11
+accuracy,AllGestureWiimoteX,Mantis,accuracy,0.7619,0.005,0,3,5
+accuracy,AllGestureWiimoteX,Mantis,accuracy,0.9624,0.0015,0,3,11
+accuracy,AllGestureWiimoteY,Mantis,accuracy,0.7948,0.0097,0,4,5
+accuracy,AllGestureWiimoteY,Mantis,accuracy,0.8516,0.0111,0,4,11
+accuracy,AllGestureWiimoteZ,Mantis,accuracy,0.73,0.01,0,5,5
+accuracy,AllGestureWiimoteZ,Mantis,accuracy,0.8087,0.0189,0,5,11
+accuracy,ArrowHead,Mantis,accuracy,0.821,0.0389,0,6,5
+accuracy,ArrowHead,Mantis,accuracy,0.7397,0.0362,0,6,11
+accuracy,BME,Mantis,accuracy,0.9956,0.0077,0,7,5
+accuracy,BME,Mantis,accuracy,0.9404,0.0072,0,7,11
+accuracy,Beef,Mantis,accuracy,0.7,0.0333,0,8,5
+accuracy,Beef,Mantis,accuracy,0.9333,0.0289,0,8,11
+accuracy,BeetleFly,Mantis,accuracy,0.8833,0.0764,0,9,5
+accuracy,BeetleFly,Mantis,accuracy,0.7662,0.0178,0,9,11
+accuracy,BirdChicken,Mantis,accuracy,0.9,0.05,0,10,5
+accuracy,BirdChicken,Mantis,accuracy,0.9552,0.001,0,10,11
+accuracy,CBF,Mantis,accuracy,0.9848,0.0074,0,11,5
+accuracy,CBF,Mantis,accuracy,0.5996,0.0209,0,11,11
+accuracy,Car,Mantis,accuracy,0.8722,0.0096,0,12,5
+accuracy,Car,Mantis,accuracy,0.8339,0.0139,0,12,11
+accuracy,Chinatown,Mantis,accuracy,0.9718,0.0017,0,13,5
+accuracy,Chinatown,Mantis,accuracy,0.4827,0.03,0,13,11
+accuracy,ChlorineConcentration,Mantis,accuracy,0.6971,0.0112,0,14,5
+accuracy,ChlorineConcentration,Mantis,accuracy,0.9531,0.0038,0,14,11
+accuracy,CinCECGTorso,Mantis,accuracy,0.7814,0.0173,0,15,5
+accuracy,CinCECGTorso,Mantis,accuracy,0.9068,0.0174,0,15,11
+accuracy,Coffee,Mantis,accuracy,1.0,0.0,0,16,5
+accuracy,Coffee,Mantis,accuracy,0.9086,0.0063,0,16,11
+accuracy,Computers,Mantis,accuracy,0.7813,0.0162,0,17,5
+accuracy,Computers,Mantis,accuracy,0.9216,0.0555,0,17,11
+accuracy,CricketX,Mantis,accuracy,0.7966,0.009,0,18,5
+accuracy,CricketX,Mantis,accuracy,0.9642,0.0024,0,18,11
+accuracy,CricketY,Mantis,accuracy,0.806,0.0141,0,19,5
+accuracy,CricketY,Mantis,accuracy,0.8889,0.0509,0,19,11
+accuracy,CricketZ,Mantis,accuracy,0.8043,0.0415,0,20,5
+accuracy,CricketZ,Mantis,accuracy,0.8162,0.0128,0,20,11
+accuracy,Crop,Mantis,accuracy,0.7444,0.0039,0,21,5
+accuracy,Crop,Mantis,accuracy,0.3214,0.0059,0,21,11
+accuracy,DiatomSizeReduction,Mantis,accuracy,0.9684,0.0136,0,22,5
+accuracy,DiatomSizeReduction,Mantis,accuracy,0.76,0.0529,0,22,11
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis,accuracy,0.7698,0.0259,0,23,5
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis,accuracy,1.0,0.0,0,23,11
+accuracy,DistalPhalanxOutlineCorrect,Mantis,accuracy,0.7717,0.0158,0,24,5
+accuracy,DistalPhalanxOutlineCorrect,Mantis,accuracy,0.9944,0.0096,0,24,11
+accuracy,DistalPhalanxTW,Mantis,accuracy,0.6954,0.0231,0,25,5
+accuracy,DistalPhalanxTW,Mantis,accuracy,0.8455,0.0246,0,25,11
+accuracy,DodgerLoopDay,Mantis,accuracy,0.625,0.025,0,26,5
+accuracy,DodgerLoopDay,Mantis,accuracy,0.8763,0.0209,0,26,11
+accuracy,DodgerLoopGame,Mantis,accuracy,0.8841,0.0145,0,27,5
+accuracy,DodgerLoopGame,Mantis,accuracy,0.7772,0.0314,0,27,11
+accuracy,DodgerLoopWeekend,Mantis,accuracy,0.9783,0.0,0,28,5
+accuracy,DodgerLoopWeekend,Mantis,accuracy,0.4916,0.0111,0,28,11
+accuracy,ECG200,Mantis,accuracy,0.8567,0.0058,0,29,5
+accuracy,ECG200,Mantis,accuracy,0.7133,0.0231,0,29,11
+accuracy,ECG5000,Mantis,accuracy,0.9335,0.0081,0,30,5
+accuracy,ECG5000,Mantis,accuracy,0.4782,0.0041,0,30,11
+accuracy,ECGFiveDays,Mantis,accuracy,0.9148,0.0292,0,31,5
+accuracy,ECGFiveDays,Mantis,accuracy,0.92,0.0,0,31,11
+accuracy,Earthquakes,Mantis,accuracy,0.7482,0.,0,32,5
+accuracy,Earthquakes,Mantis,accuracy,1.0,0.0,0,32,11
+accuracy,ElectricDevices,Mantis,accuracy,0.7454,0.0049,0,33,5
+accuracy,ElectricDevices,Mantis,accuracy,0.8906,0.0059,0,33,11
+accuracy,FaceAll,Mantis,accuracy,0.8308,0.0031,0,34,5
+accuracy,FaceAll,Mantis,accuracy,0.7929,0.0031,0,34,11
+accuracy,FaceFour,Mantis,accuracy,0.9773,0.0114,0,35,5
+accuracy,FaceFour,Mantis,accuracy,0.9978,0.0038,0,35,11
+accuracy,FacesUCR,Mantis,accuracy,0.9148,0.0071,0,36,5
+accuracy,FacesUCR,Mantis,accuracy,0.8142,0.0113,0,36,11
+accuracy,FiftyWords,Mantis,accuracy,0.8139,0.0083,0,37,5
+accuracy,FiftyWords,Mantis,accuracy,0.8996,0.0112,0,37,11
+accuracy,Fish,Mantis,accuracy,0.9714,0.0099,0,38,5
+accuracy,Fish,Mantis,accuracy,0.9667,0.0078,0,38,11
+accuracy,FordA,Mantis,accuracy,0.9343,0.0046,0,39,5
+accuracy,FordA,Mantis,accuracy,0.9712,0.0042,0,39,11
+accuracy,FordB,Mantis,accuracy,0.7979,0.0135,0,40,5
+accuracy,FordB,Mantis,accuracy,0.9889,0.0017,0,40,11
+accuracy,FreezerRegularTrain,Mantis,accuracy,0.9927,0.0041,0,41,5
+accuracy,FreezerRegularTrain,Mantis,accuracy,0.9944,0.0038,0,41,11
+accuracy,FreezerSmallTrain,Mantis,accuracy,0.9667,0.0098,0,42,5
+accuracy,FreezerSmallTrain,Mantis,accuracy,0.9664,0.0067,0,42,11
+accuracy,Fungi,Mantis,accuracy,0.7778,0.0031,0,43,5
+accuracy,ACSF1,Mantis-RF,accuracy,0.6133,0.0208,2,1,1
+accuracy,ACSF1,Mantis-Head,accuracy,0.62,0.0265,2,1,2
+accuracy,ACSF1,Mantis-Scratch,accuracy,0.6433,0.0451,2,1,3
+accuracy,ACSF1,Mantis-Full,accuracy,0.7467,0.0416,2,1,4
+accuracy,Adiac,Mantis-RF,accuracy,0.7332,0.0039,2,2,1
+accuracy,Adiac,Mantis-Head,accuracy,0.5857,0.0051,2,2,2
+accuracy,Adiac,Mantis-Scratch,accuracy,0.7613,0.0129,2,2,3
+accuracy,Adiac,Mantis-Full,accuracy,0.8295,0.0082,2,2,4
+accuracy,AllGestureWiimoteX,Mantis-RF,accuracy,0.6705,0.0044,2,3,1
+accuracy,AllGestureWiimoteX,Mantis-Head,accuracy,0.551,0.0008,2,3,2
+accuracy,AllGestureWiimoteX,Mantis-Scratch,accuracy,0.6962,0.0058,2,3,3
+accuracy,AllGestureWiimoteX,Mantis-Full,accuracy,0.8029,0.0065,2,3,4
+accuracy,AllGestureWiimoteY,Mantis-RF,accuracy,0.6671,0.0057,2,4,1
+accuracy,AllGestureWiimoteY,Mantis-Head,accuracy,0.5614,0.0052,2,4,2
+accuracy,AllGestureWiimoteY,Mantis-Scratch,accuracy,0.739,0.0008,2,4,3
+accuracy,AllGestureWiimoteY,Mantis-Full,accuracy,0.8129,0.0089,2,4,4
+accuracy,AllGestureWiimoteZ,Mantis-RF,accuracy,0.6695,0.0033,2,5,1
+accuracy,AllGestureWiimoteZ,Mantis-Head,accuracy,0.5467,0.0008,2,5,2
+accuracy,AllGestureWiimoteZ,Mantis-Scratch,accuracy,0.6367,0.0246,2,5,3
+accuracy,AllGestureWiimoteZ,Mantis-Full,accuracy,0.7481,0.0095,2,5,4
+accuracy,ArrowHead,Mantis-RF,accuracy,0.7105,0.0175,2,6,1
+accuracy,ArrowHead,Mantis-Head,accuracy,0.7276,0.0119,2,6,2
+accuracy,ArrowHead,Mantis-Scratch,accuracy,0.8,0.0286,2,6,3
+accuracy,ArrowHead,Mantis-Full,accuracy,0.8229,0.0297,2,6,4
+accuracy,BME,Mantis-RF,accuracy,0.9311,0.0038,2,7,1
+accuracy,BME,Mantis-Head,accuracy,0.8756,0.0139,2,7,2
+accuracy,BME,Mantis-Scratch,accuracy,0.9933,0.0,2,7,3
+accuracy,BME,Mantis-Full,accuracy,1.0,0.0,2,7,4
+accuracy,Beef,Mantis-RF,accuracy,0.6556,0.0509,2,8,1
+accuracy,Beef,Mantis-Head,accuracy,0.6222,0.0385,2,8,2
+accuracy,Beef,Mantis-Scratch,accuracy,0.8222,0.0694,2,8,3
+accuracy,Beef,Mantis-Full,accuracy,0.8667,0.0,2,8,4
+accuracy,BeetleFly,Mantis-RF,accuracy,0.85,0.05,2,9,1
+accuracy,BeetleFly,Mantis-Head,accuracy,0.9,0.0,2,9,2
+accuracy,BeetleFly,Mantis-Scratch,accuracy,0.7,0.0,2,9,3
+accuracy,BeetleFly,Mantis-Full,accuracy,0.9,0.05,2,9,4
+accuracy,BirdChicken,Mantis-RF,accuracy,1.0,0.0,2,10,1
+accuracy,BirdChicken,Mantis-Head,accuracy,1.0,0.0,2,10,2
+accuracy,BirdChicken,Mantis-Scratch,accuracy,0.9,0.0,2,10,3
+accuracy,BirdChicken,Mantis-Full,accuracy,1.0,0.0,2,10,4
+accuracy,CBF,Mantis-RF,accuracy,0.993,0.0013,2,11,1
+accuracy,CBF,Mantis-Head,accuracy,0.9941,0.0006,2,11,2
+accuracy,CBF,Mantis-Scratch,accuracy,0.9922,0.008,2,11,3
+accuracy,CBF,Mantis-Full,accuracy,0.9996,0.0006,2,11,4
+accuracy,Car,Mantis-RF,accuracy,0.7722,0.0419,2,12,1
+accuracy,Car,Mantis-Head,accuracy,0.7778,0.0192,2,12,2
+accuracy,Car,Mantis-Scratch,accuracy,0.8,0.0,2,12,3
+accuracy,Car,Mantis-Full,accuracy,0.9,0.0289,2,12,4
+accuracy,Chinatown,Mantis-RF,accuracy,0.8737,0.0168,2,13,1
+accuracy,Chinatown,Mantis-Head,accuracy,0.8785,0.0094,2,13,2
+accuracy,Chinatown,Mantis-Scratch,accuracy,0.964,0.0017,2,13,3
+accuracy,Chinatown,Mantis-Full,accuracy,0.9776,0.0067,2,13,4
+accuracy,ChlorineConcentration,Mantis-RF,accuracy,0.6806,0.0019,2,14,1
+accuracy,ChlorineConcentration,Mantis-Head,accuracy,0.5704,0.0011,2,14,2
+accuracy,ChlorineConcentration,Mantis-Scratch,accuracy,0.753,0.0056,2,14,3
+accuracy,ChlorineConcentration,Mantis-Full,accuracy,0.7788,0.0072,2,14,4
+accuracy,CinCECGTorso,Mantis-RF,accuracy,0.6611,0.0036,2,15,1
+accuracy,CinCECGTorso,Mantis-Head,accuracy,0.7179,0.0048,2,15,2
+accuracy,CinCECGTorso,Mantis-Scratch,accuracy,0.7901,0.0198,2,15,3
+accuracy,CinCECGTorso,Mantis-Full,accuracy,0.856,0.0077,2,15,4
+accuracy,Coffee,Mantis-RF,accuracy,0.9524,0.0206,2,16,1
+accuracy,Coffee,Mantis-Head,accuracy,0.9643,0.0,2,16,2
+accuracy,Coffee,Mantis-Scratch,accuracy,1.0,0.0,2,16,3
+accuracy,Coffee,Mantis-Full,accuracy,1.0,0.0,2,16,4
+accuracy,Computers,Mantis-RF,accuracy,0.7373,0.0092,2,17,1
+accuracy,Computers,Mantis-Head,accuracy,0.728,0.0069,2,17,2
+accuracy,Computers,Mantis-Scratch,accuracy,0.7533,0.0162,2,17,3
+accuracy,Computers,Mantis-Full,accuracy,0.7773,0.0046,2,17,4
+accuracy,CricketX,Mantis-RF,accuracy,0.7368,0.0171,2,18,1
+accuracy,CricketX,Mantis-Head,accuracy,0.6966,0.003,2,18,2
+accuracy,CricketX,Mantis-Scratch,accuracy,0.7812,0.0146,2,18,3
+accuracy,CricketX,Mantis-Full,accuracy,0.8282,0.0133,2,18,4
+accuracy,CricketY,Mantis-RF,accuracy,0.7504,0.0065,2,19,1
+accuracy,CricketY,Mantis-Head,accuracy,0.7051,0.0026,2,19,2
+accuracy,CricketY,Mantis-Scratch,accuracy,0.8,0.0026,2,19,3
+accuracy,CricketY,Mantis-Full,accuracy,0.8385,0.0051,2,19,4
+accuracy,CricketZ,Mantis-RF,accuracy,0.7900,0.0039,2,20,1
+accuracy,CricketZ,Mantis-Head,accuracy,0.7333,0.0077,2,20,2
+accuracy,CricketZ,Mantis-Scratch,accuracy,0.7991,0.0097,2,20,3
+accuracy,CricketZ,Mantis-Full,accuracy,0.8521,0.0082,2,20,4
+accuracy,Crop,Mantis-RF,accuracy,0.6756,0.0018,2,21,1
+accuracy,Crop,Mantis-Head,accuracy,0.6745,0.0001,2,21,2
+accuracy,Crop,Mantis-Scratch,accuracy,0.7566,0.0063,2,21,3
+accuracy,Crop,Mantis-Full,accuracy,0.7702,0.0005,2,21,4
+accuracy,DiatomSizeReduction,Mantis-RF,accuracy,0.8845,0.0019,2,22,1
+accuracy,DiatomSizeReduction,Mantis-Head,accuracy,0.8911,0.0191,2,22,2
+accuracy,DiatomSizeReduction,Mantis-Scratch,accuracy,0.9031,0.0147,2,22,3
+accuracy,DiatomSizeReduction,Mantis-Full,accuracy,0.9346,0.0033,2,22,4
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-RF,accuracy,0.789,0.015,2,23,1
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-Head,accuracy,0.8106,0.0042,2,23,2
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-Scratch,accuracy,0.7506,0.011,2,23,3
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis-Full,accuracy,0.7938,0.0042,2,23,4
+accuracy,DistalPhalanxOutlineCorrect,Mantis-RF,accuracy,0.75,0.0126,2,24,1
+accuracy,DistalPhalanxOutlineCorrect,Mantis-Head,accuracy,0.7729,0.0055,2,24,2
+accuracy,DistalPhalanxOutlineCorrect,Mantis-Scratch,accuracy,0.8056,0.0091,2,24,3
+accuracy,DistalPhalanxOutlineCorrect,Mantis-Full,accuracy,0.7935,0.0036,2,24,4
+accuracy,DistalPhalanxTW,Mantis-RF,accuracy,0.6859,0.011,2,25,1
+accuracy,DistalPhalanxTW,Mantis-Head,accuracy,0.6763,0.0125,2,25,2
+accuracy,DistalPhalanxTW,Mantis-Scratch,accuracy,0.7266,0.0125,2,25,3
+accuracy,DistalPhalanxTW,Mantis-Full,accuracy,0.7338,0.019,2,25,4
+accuracy,DodgerLoopDay,Mantis-RF,accuracy,0.55,0.0217,2,26,1
+accuracy,DodgerLoopDay,Mantis-Head,accuracy,0.5292,0.026,2,26,2
+accuracy,DodgerLoopDay,Mantis-Scratch,accuracy,0.6208,0.0191,2,26,3
+accuracy,DodgerLoopDay,Mantis-Full,accuracy,0.6375,0.0125,2,26,4
+accuracy,DodgerLoopGame,Mantis-RF,accuracy,0.7585,0.0221,2,27,1
+accuracy,DodgerLoopGame,Mantis-Head,accuracy,0.7222,0.0111,2,27,2
+accuracy,DodgerLoopGame,Mantis-Scratch,accuracy,0.843,0.0084,2,27,3
+accuracy,DodgerLoopGame,Mantis-Full,accuracy,0.9541,0.0042,2,27,4
+accuracy,DodgerLoopWeekend,Mantis-RF,accuracy,0.9517,0.0084,2,28,1
+accuracy,DodgerLoopWeekend,Mantis-Head,accuracy,0.9783,0.0126,2,28,2
+accuracy,DodgerLoopWeekend,Mantis-Scratch,accuracy,0.9855,0.0,2,28,3
+accuracy,DodgerLoopWeekend,Mantis-Full,accuracy,0.9855,0.0,2,28,4
+accuracy,ECG200,Mantis-RF,accuracy,0.82,0.01,2,29,1
+accuracy,ECG200,Mantis-Head,accuracy,0.8033,0.0252,2,29,2
+accuracy,ECG200,Mantis-Scratch,accuracy,0.8467,0.0115,2,29,3
+accuracy,ECG200,Mantis-Full,accuracy,0.9133,0.0115,2,29,4
+accuracy,ECG5000,Mantis-RF,accuracy,0.9211,0.001,2,30,1
+accuracy,ECG5000,Mantis-Head,accuracy,0.9185,0.0038,2,30,2
+accuracy,ECG5000,Mantis-Scratch,accuracy,0.9411,0.0008,2,30,3
+accuracy,ECG5000,Mantis-Full,accuracy,0.9459,0.0014,2,30,4
+accuracy,ECGFiveDays,Mantis-RF,accuracy,0.909,0.0218,2,31,1
+accuracy,ECGFiveDays,Mantis-Head,accuracy,0.8668,0.0401,2,31,2
+accuracy,ECGFiveDays,Mantis-Scratch,accuracy,0.7855,0.0169,2,31,3
+accuracy,ECGFiveDays,Mantis-Full,accuracy,0.9694,0.014,2,31,4
+accuracy,EOGHorizontalSignal,Mantis-RF,accuracy,0.5875,0.0089,2,32,1
+accuracy,EOGHorizontalSignal,Mantis-Head,accuracy,0.5783,0.0064,2,32,2
+accuracy,EOGHorizontalSignal,Mantis-Scratch,accuracy,0.6703,0.0194,2,32,3
+accuracy,EOGHorizontalSignal,Mantis-Full,accuracy,0.6934,0.012,2,32,4
+accuracy,EOGVerticalSignal,Mantis-RF,accuracy,0.4751,0.0,2,33,1
+accuracy,EOGVerticalSignal,Mantis-Head,accuracy,0.4825,0.0105,2,33,2
+accuracy,EOGVerticalSignal,Mantis-Scratch,accuracy,0.5396,0.0058,2,33,3
+accuracy,EOGVerticalSignal,Mantis-Full,accuracy,0.5773,0.0073,2,33,4
+accuracy,Earthquakes,Mantis-RF,accuracy,0.7482,0.0,2,34,1
+accuracy,Earthquakes,Mantis-Head,accuracy,0.7506,0.0042,2,34,2
+accuracy,Earthquakes,Mantis-Scratch,accuracy,0.7866,0.015,2,34,3
+accuracy,Earthquakes,Mantis-Full,accuracy,0.7602,0.015,2,34,4
+accuracy,ElectricDevices,Mantis-RF,accuracy,0.7226,0.0026,2,35,1
+accuracy,ElectricDevices,Mantis-Head,accuracy,0.7064,0.001,2,35,2
+accuracy,ElectricDevices,Mantis-Scratch,accuracy,0.7448,0.0044,2,35,3
+accuracy,ElectricDevices,Mantis-Full,accuracy,0.7643,0.005,2,35,4
+accuracy,EthanolLevel,Mantis-RF,accuracy,0.2993,0.011,2,36,1
+accuracy,EthanolLevel,Mantis-Head,accuracy,0.324,0.006,2,36,2
+accuracy,EthanolLevel,Mantis-Scratch,accuracy,0.6987,0.0194,2,36,3
+accuracy,EthanolLevel,Mantis-Full,accuracy,0.7393,0.1052,2,36,4
+accuracy,FaceAll,Mantis-RF,accuracy,0.7815,0.0074,2,37,1
+accuracy,FaceAll,Mantis-Head,accuracy,0.8053,0.0062,2,37,2
+accuracy,FaceAll,Mantis-Scratch,accuracy,0.9542,0.0036,2,37,3
+accuracy,FaceAll,Mantis-Full,accuracy,0.8554,0.0343,2,37,4
+accuracy,FaceFour,Mantis-RF,accuracy,0.9508,0.0066,2,38,1
+accuracy,FaceFour,Mantis-Head,accuracy,0.9508,0.0174,2,38,2
+accuracy,FaceFour,Mantis-Scratch,accuracy,0.8485,0.0347,2,38,3
+accuracy,FaceFour,Mantis-Full,accuracy,0.9811,0.0066,2,38,4
+accuracy,FacesUCR,Mantis-RF,accuracy,0.8354,0.0054,2,39,1
+accuracy,FacesUCR,Mantis-Head,accuracy,0.8437,0.0104,2,39,2
+accuracy,FacesUCR,Mantis-Scratch,accuracy,0.8816,0.0054,2,39,3
+accuracy,FacesUCR,Mantis-Full,accuracy,0.942,0.0061,2,39,4
+accuracy,FiftyWords,Mantis-RF,accuracy,0.6462,0.0096,2,40,1
+accuracy,FiftyWords,Mantis-Head,accuracy,0.7026,0.0178,2,40,2
+accuracy,FiftyWords,Mantis-Scratch,accuracy,0.8176,0.0038,2,40,3
+accuracy,FiftyWords,Mantis-Full,accuracy,0.863,0.0013,2,40,4
+accuracy,Fish,Mantis-RF,accuracy,0.9333,0.0066,2,41,1
+accuracy,Fish,Mantis-Head,accuracy,0.9333,0.0033,2,41,2
+accuracy,Fish,Mantis-Scratch,accuracy,0.939,0.0087,2,41,3
+accuracy,Fish,Mantis-Full,accuracy,0.979,0.0087,2,41,4
+accuracy,FordA,Mantis-RF,accuracy,0.8581,0.0048,2,42,1
+accuracy,FordA,Mantis-Head,accuracy,0.8881,0.0004,2,42,2
+accuracy,FordA,Mantis-Scratch,accuracy,0.9301,0.0031,2,42,3
+accuracy,FordA,Mantis-Full,accuracy,0.9477,0.0013,2,42,4
+accuracy,FordB,Mantis-RF,accuracy,0.7305,0.0031,2,43,1
+accuracy,FordB,Mantis-Head,accuracy,0.7704,0.0033,2,43,2
+accuracy,FordB,Mantis-Scratch,accuracy,0.8115,0.0036,2,43,3
+accuracy,FordB,Mantis-Full,accuracy,0.8477,0.004,2,43,4
+accuracy,FreezerRegularTrain,Mantis-RF,accuracy,0.9374,0.0043,2,44,1
+accuracy,FreezerRegularTrain,Mantis-Head,accuracy,0.8767,0.0088,2,44,2
+accuracy,FreezerRegularTrain,Mantis-Scratch,accuracy,0.9971,0.0004,2,44,3
+accuracy,FreezerRegularTrain,Mantis-Full,accuracy,0.9967,0.0014,2,44,4
+accuracy,FreezerSmallTrain,Mantis-RF,accuracy,0.7942,0.0059,2,45,1
+accuracy,FreezerSmallTrain,Mantis-Head,accuracy,0.8092,0.0041,2,45,2
+accuracy,FreezerSmallTrain,Mantis-Scratch,accuracy,0.8601,0.073,2,45,3
+accuracy,FreezerSmallTrain,Mantis-Full,accuracy,0.9648,0.0219,2,45,4
+accuracy,Fungi,Mantis-RF,accuracy,0.8262,0.0164,2,46,1
+accuracy,Fungi,Mantis-Head,accuracy,0.8477,0.0224,2,46,2
+accuracy,Fungi,Mantis-Scratch,accuracy,0.948,0.0164,2,46,3
+accuracy,Fungi,Mantis-Full,accuracy,0.9892,0.0108,2,46,4
+accuracy,GestureMidAirD1,Mantis-RF,accuracy,0.659,0.0044,2,47,1
+accuracy,GestureMidAirD1,Mantis-Head,accuracy,0.6026,0.0044,2,47,2
+accuracy,GestureMidAirD1,Mantis-Scratch,accuracy,0.7103,0.0089,2,47,3
+accuracy,GestureMidAirD1,Mantis-Full,accuracy,0.7949,0.0044,2,47,4
+accuracy,GestureMidAirD2,Mantis-RF,accuracy,0.6154,0.0077,2,48,1
+accuracy,GestureMidAirD2,Mantis-Head,accuracy,0.5282,0.0118,2,48,2
+accuracy,GestureMidAirD2,Mantis-Scratch,accuracy,0.6974,0.0222,2,48,3
+accuracy,GestureMidAirD2,Mantis-Full,accuracy,0.7205,0.0247,2,48,4
+accuracy,GestureMidAirD3,Mantis-RF,accuracy,0.3282,0.016,2,49,1
+accuracy,GestureMidAirD3,Mantis-Head,accuracy,0.3154,0.0266,2,49,2
+accuracy,GestureMidAirD3,Mantis-Scratch,accuracy,0.4359,0.016,2,49,3
+accuracy,GestureMidAirD3,Mantis-Full,accuracy,0.4795,0.0178,2,49,4
+accuracy,GesturePebbleZ1,Mantis-RF,accuracy,0.9283,0.0034,2,50,1
+accuracy,GesturePebbleZ1,Mantis-Head,accuracy,0.9186,0.0116,2,50,2
+accuracy,GesturePebbleZ1,Mantis-Scratch,accuracy,0.9322,0.0089,2,50,3
+accuracy,GesturePebbleZ1,Mantis-Full,accuracy,0.9574,0.0146,2,50,4
+accuracy,GesturePebbleZ2,Mantis-RF,accuracy,0.9219,0.0256,2,51,1
+accuracy,GesturePebbleZ2,Mantis-Head,accuracy,0.9515,0.0097,2,51,2
+accuracy,GesturePebbleZ2,Mantis-Scratch,accuracy,0.8882,0.0193,2,51,3
+accuracy,GesturePebbleZ2,Mantis-Full,accuracy,0.9451,0.0159,2,51,4
+accuracy,GunPoint,Mantis-RF,accuracy,0.98,0.0067,2,52,1
+accuracy,GunPoint,Mantis-Head,accuracy,0.9822,0.0038,2,52,2
+accuracy,GunPoint,Mantis-Scratch,accuracy,0.9867,0.0133,2,52,3
+accuracy,GunPoint,Mantis-Full,accuracy,0.9978,0.0038,2,52,4
+accuracy,GunPointAgeSpan,Mantis-RF,accuracy,0.9905,0.0,2,53,1
+accuracy,GunPointAgeSpan,Mantis-Head,accuracy,0.9821,0.0037,2,53,2
+accuracy,GunPointAgeSpan,Mantis-Scratch,accuracy,0.9947,0.0048,2,53,3
+accuracy,GunPointAgeSpan,Mantis-Full,accuracy,0.9989,0.0018,2,53,4
+accuracy,GunPointMaleVersusFemale,Mantis-RF,accuracy,0.9958,0.0018,2,54,1
+accuracy,GunPointMaleVersusFemale,Mantis-Head,accuracy,0.9863,0.0048,2,54,2
+accuracy,GunPointMaleVersusFemale,Mantis-Scratch,accuracy,0.9947,0.0018,2,54,3
+accuracy,GunPointMaleVersusFemale,Mantis-Full,accuracy,1.0,0.0,2,54,4
+accuracy,GunPointOldVersusYoung,Mantis-RF,accuracy,0.9968,0.0,2,55,1
+accuracy,GunPointOldVersusYoung,Mantis-Head,accuracy,0.9968,0.0,2,55,2
+accuracy,GunPointOldVersusYoung,Mantis-Scratch,accuracy,0.9968,0.0,2,55,3
+accuracy,GunPointOldVersusYoung,Mantis-Full,accuracy,1.0,0.0,2,55,4
+accuracy,Ham,Mantis-RF,accuracy,0.673,0.0145,2,56,1
+accuracy,Ham,Mantis-Head,accuracy,0.7143,0.0252,2,56,2
+accuracy,Ham,Mantis-Scratch,accuracy,0.8159,0.0198,2,56,3
+accuracy,Ham,Mantis-Full,accuracy,0.7016,0.0198,2,56,4
+accuracy,HandOutlines,Mantis-RF,accuracy,0.9162,0.0072,2,57,1
+accuracy,HandOutlines,Mantis-Head,accuracy,0.8928,0.0068,2,57,2
+accuracy,HandOutlines,Mantis-Scratch,accuracy,0.945,0.0056,2,57,3
+accuracy,HandOutlines,Mantis-Full,accuracy,0.955,0.0016,2,57,4
+accuracy,Haptics,Mantis-RF,accuracy,0.4968,0.0032,2,58,1
+accuracy,Haptics,Mantis-Head,accuracy,0.4697,0.0131,2,58,2
+accuracy,Haptics,Mantis-Scratch,accuracy,0.5032,0.0065,2,58,3
+accuracy,Haptics,Mantis-Full,accuracy,0.5747,0.0234,2,58,4
+accuracy,Herring,Mantis-RF,accuracy,0.6667,0.0239,2,59,1
+accuracy,Herring,Mantis-Head,accuracy,0.7188,0.0271,2,59,2
+accuracy,Herring,Mantis-Scratch,accuracy,0.6458,0.0239,2,59,3
+accuracy,Herring,Mantis-Full,accuracy,0.724,0.009,2,59,4
+accuracy,HouseTwenty,Mantis-RF,accuracy,0.9412,0.0,2,60,1
+accuracy,HouseTwenty,Mantis-Head,accuracy,0.9496,0.0084,2,60,2
+accuracy,HouseTwenty,Mantis-Scratch,accuracy,0.9356,0.0049,2,60,3
+accuracy,HouseTwenty,Mantis-Full,accuracy,0.9776,0.0097,2,60,4
+accuracy,InlineSkate,Mantis-RF,accuracy,0.363,0.0136,2,61,1
+accuracy,InlineSkate,Mantis-Head,accuracy,0.3182,0.0101,2,61,2
+accuracy,InlineSkate,Mantis-Scratch,accuracy,0.403,0.0211,2,61,3
+accuracy,InlineSkate,Mantis-Full,accuracy,0.4515,0.0132,2,61,4
+accuracy,InsectEPGRegularTrain,Mantis-RF,accuracy,1.0,0.0,2,62,1
+accuracy,InsectEPGRegularTrain,Mantis-Head,accuracy,1.0,0.0,2,62,2
+accuracy,InsectEPGRegularTrain,Mantis-Scratch,accuracy,0.992,0.0,2,62,3
+accuracy,InsectEPGRegularTrain,Mantis-Full,accuracy,1.0,0.0,2,62,4
+accuracy,InsectEPGSmallTrain,Mantis-RF,accuracy,1.0,0.0,2,63,1
+accuracy,InsectEPGSmallTrain,Mantis-Head,accuracy,1.0,0.0,2,63,2
+accuracy,InsectEPGSmallTrain,Mantis-Scratch,accuracy,0.9665,0.0061,2,63,3
+accuracy,InsectEPGSmallTrain,Mantis-Full,accuracy,1.0,0.0,2,63,4
+accuracy,InsectWingbeatSound,Mantis-RF,accuracy,0.519,0.0044,2,64,1
+accuracy,InsectWingbeatSound,Mantis-Head,accuracy,0.4791,0.0099,2,64,2
+accuracy,InsectWingbeatSound,Mantis-Scratch,accuracy,0.6222,0.0066,2,64,3
+accuracy,InsectWingbeatSound,Mantis-Full,accuracy,0.6246,0.0068,2,64,4
+accuracy,Blink,Mantis-RF,accuracy,1.0,0.0,2,65,1
+accuracy,Blink,Mantis-Head,accuracy,0.9963,0.0046,2,65,2
+accuracy,Blink,Mantis-Scratch,accuracy,0.9896,0.0056,2,65,3
+accuracy,Blink,Mantis-Full,accuracy,1.0,0.0,2,65,4
+accuracy,MotionSenseHAR,Mantis-RF,accuracy,1.0,0.0,2,66,1
+accuracy,MotionSenseHAR,Mantis-Head,accuracy,0.9962,0.0,2,66,2
+accuracy,MotionSenseHAR,Mantis-Scratch,accuracy,0.9887,0.0,2,66,3
+accuracy,MotionSenseHAR,Mantis-Full,accuracy,1.0,0.0,2,66,4
+accuracy,SharePriceIncrease,Mantis-RF,accuracy,0.6874,0.0,2,67,1
+accuracy,SharePriceIncrease,Mantis-Head,accuracy,0.6746,0.0012,2,67,2
+accuracy,SharePriceIncrease,Mantis-Scratch,accuracy,0.6891,0.0039,2,67,3
+accuracy,SharePriceIncrease,Mantis-Full,accuracy,0.6874,0.0027,2,67,4
+accuracy,ArticularyWordRecognition,Mantis-RF,accuracy,0.9933,0.0033,2,68,1
+accuracy,ArticularyWordRecognition,Mantis-Head,accuracy,0.9933,0.0,2,68,2
+accuracy,ArticularyWordRecognition,Mantis-Scratch,accuracy,0.9967,0.0,2,68,3
+accuracy,ArticularyWordRecognition,Mantis-Full,accuracy,0.9933,0.0,2,68,4
+accuracy,BasicMotions,Mantis-RF,accuracy,1.0,0.0,2,69,1
+accuracy,BasicMotions,Mantis-Head,accuracy,1.0,0.0,2,69,2
+accuracy,BasicMotions,Mantis-Scratch,accuracy,1.0,0.0,2,69,3
+accuracy,BasicMotions,Mantis-Full,accuracy,1.0,0.0,2,69,4
+accuracy,CharacterTrajectories,Mantis-RF,accuracy,0.9396,0.0008,2,70,1
+accuracy,CharacterTrajectories,Mantis-Head,accuracy,0.9673,0.0021,2,70,2
+accuracy,CharacterTrajectories,Mantis-Scratch,accuracy,0.9838,0.0029,2,70,3
+accuracy,CharacterTrajectories,Mantis-Full,accuracy,0.9928,0.0004,2,70,4
+accuracy,Cricket,Mantis-RF,accuracy,1.0,0.0,2,71,1
+accuracy,Cricket,Mantis-Head,accuracy,0.9907,0.008,2,71,2
+accuracy,Cricket,Mantis-Scratch,accuracy,0.9861,0.0,2,71,3
+accuracy,Cricket,Mantis-Full,accuracy,1.0,0.0,2,71,4
+accuracy,ERing,Mantis-RF,accuracy,0.9407,0.0098,2,72,1
+accuracy,ERing,Mantis-Head,accuracy,0.9296,0.0192,2,72,2
+accuracy,ERing,Mantis-Scratch,accuracy,0.9432,0.0077,2,72,3
+accuracy,ERing,Mantis-Full,accuracy,0.9926,0.0074,2,72,4
+accuracy,EigenWorms,Mantis-RF,accuracy,0.7532,0.0159,2,73,1
+accuracy,EigenWorms,Mantis-Head,accuracy,0.7379,0.0233,2,73,2
+accuracy,EigenWorms,Mantis-Scratch,accuracy,0.7023,0.0202,2,73,3
+accuracy,EigenWorms,Mantis-Full,accuracy,0.8372,0.0044,2,73,4
+accuracy,Epilepsy,Mantis-RF,accuracy,0.9952,0.0042,2,74,1
+accuracy,Epilepsy,Mantis-Head,accuracy,0.9928,0.0,2,74,2
+accuracy,Epilepsy,Mantis-Scratch,accuracy,1.0,0.0,2,74,3
+accuracy,Epilepsy,Mantis-Full,accuracy,1.0,0.0,2,74,4
+accuracy,EthanolConcentration,Mantis-RF,accuracy,0.27,0.0101,2,75,1
+accuracy,EthanolConcentration,Mantis-Head,accuracy,0.3156,0.0076,2,75,2
+accuracy,EthanolConcentration,Mantis-Scratch,accuracy,0.3321,0.0154,2,75,3
+accuracy,EthanolConcentration,Mantis-Full,accuracy,0.4208,0.0195,2,75,4
+accuracy,HandMovementDirection,Mantis-RF,accuracy,0.2117,0.0413,2,76,1
+accuracy,HandMovementDirection,Mantis-Head,accuracy,0.3829,0.0639,2,76,2
+accuracy,HandMovementDirection,Mantis-Scratch,accuracy,0.5135,0.0234,2,76,3
+accuracy,HandMovementDirection,Mantis-Full,accuracy,0.4009,0.0206,2,76,4
+accuracy,ACSF1,Mantis,performance,0.6133,,5,1,3
+accuracy,Adiac,Mantis,performance,0.7332,,5,2,3
+accuracy,AllGestureWiimoteX,Mantis,performance,0.6705,,5,3,3
+accuracy,AllGestureWiimoteY,Mantis,performance,0.6671,,5,4,3
+accuracy,AllGestureWiimoteZ,Mantis,performance,0.6695,,5,5,3
+accuracy,ArrowHead,Mantis,performance,0.7105,,5,6,3
+accuracy,BME,Mantis,performance,0.9311,,5,7,3
+accuracy,Beef,Mantis,performance,0.6556,,5,8,3
+accuracy,BeetleFly,Mantis,performance,0.85,,5,9,3
+accuracy,BirdChicken,Mantis,performance,1.0,,5,10,3
+accuracy,CBF,Mantis,performance,0.993,,5,11,3
+accuracy,Car,Mantis,performance,0.7722,,5,12,3
+accuracy,Chinatown,Mantis,performance,0.8737,,5,13,3
+accuracy,ChlorineConcentration,Mantis,performance,0.6806,,5,14,3
+accuracy,CinCECGTorso,Mantis,performance,0.6611,,5,15,3
+accuracy,Coffee,Mantis,performance,0.9524,,5,16,3
+accuracy,Computers,Mantis,performance,0.7373,,5,17,3
+accuracy,CricketX,Mantis,performance,0.7368,,5,18,3
+accuracy,CricketY,Mantis,performance,0.7504,,5,19,3
+accuracy,CricketZ,Mantis,performance,0.7906,,5,20,3
+accuracy,Crop,Mantis,performance,0.6756,,5,21,3
+accuracy,DiatomSizeReduction,Mantis,performance,0.8845,,5,22,3
+accuracy,DistalPhalanxOutlineAgeGroup,Mantis,performance,0.789,,5,23,3
+accuracy,DistalPhalanxOutlineCorrect,Mantis,performance,0.75,,5,24,3
+accuracy,DistalPhalanxTW,Mantis,performance,0.6859,,5,25,3
+accuracy,DodgerLoopDay,Mantis,performance,0.55,,5,26,3
+accuracy,DodgerLoopGame,Mantis,performance,0.7585,,5,27,3
+accuracy,DodgerLoopWeekend,Mantis,performance,0.9517,,5,28,3
+accuracy,ECG200,Mantis,performance,0.82,,5,29,3
+accuracy,ECG5000,Mantis,performance,0.9211,,5,30,3
+accuracy,ECGFiveDays,Mantis,performance,0.909,,5,31,3
+accuracy,EOGHorizontalSignal,Mantis,performance,0.5875,,5,32,3
+accuracy,EOGVerticalSignal,Mantis,performance,0.4751,,5,33,3
+accuracy,Earthquakes,Mantis,performance,0.7482,,5,34,3
+accuracy,ElectricDevices,Mantis,performance,0.7226,,5,35,3
+accuracy,EthanolLevel,Mantis,performance,0.2993,,5,36,3
+accuracy,FaceAll,Mantis,performance,0.7815,,5,37,3
+accuracy,FaceFour,Mantis,performance,0.9508,,5,38,3
+accuracy,FacesUCR,Mantis,performance,0.8354,,5,39,3
+accuracy,FiftyWords,Mantis,performance,0.6462,,5,40,3
+accuracy,Fish,Mantis,performance,0.9333,,5,41,3
+accuracy,FordA,Mantis,performance,0.8581,,5,42,3
+accuracy,FordB,Mantis,performance,0.7305,,5,43,3
+accuracy,FreezerRegularTrain,Mantis,performance,0.9374,,5,44,3
+accuracy,FreezerSmallTrain,Mantis,performance,0.7942,,5,45,3
+accuracy,Fungi,Mantis,performance,0.8262,,5,46,3
+accuracy,GestureMidAirD2,Mantis,performance,0.6154,,5,48,3
+accuracy,GestureMidAirD3,Mantis,performance,0.3282,,5,49,3
+accuracy,GesturePebbleZ1,Mantis,performance,0.9283,,5,50,3
+accuracy,GesturePebbleZ2,Mantis,performance,0.9219,,5,51,3
+accuracy,GunPoint,Mantis,performance,0.98,,5,52,3
+accuracy,GunPointAgeSpan,Mantis,performance,0.9905,,5,53,3
+accuracy,GunPointMaleVersusFemale,Mantis,performance,0.9958,,5,54,3
+accuracy,GunPointOldVersusYoung,Mantis,performance,0.9968,,5,55,3
+accuracy,Ham,Mantis,performance,0.673,,5,56,3
+accuracy,HandOutlines,Mantis,performance,0.9162,,5,57,3
+accuracy,Haptics,Mantis,performance,0.4968,,5,58,3
+accuracy,Herring,Mantis,performance,0.6667,,5,59,3
+accuracy,HouseTwenty,Mantis,performance,0.9412,,5,60,3
+accuracy,InlineSkate,Mantis,performance,0.363,,5,61,3
+accuracy,InsectWingbeatSound,Mantis,performance,0.519,,5,64,3
+accuracy,Blink,Mantis,performance,0.9993,,5,65,3
+accuracy,EMOPain,Mantis,performance,0.8385,,5,66,3
+accuracy,MotionSenseHAR,Mantis,performance,0.9975,,5,67,3
+accuracy,SharePriceIncrease,Mantis,performance,0.6836,,5,68,3
+accuracy,ArticularyWordRecognition,Mantis,performance,0.993,,5,69,3
+accuracy,CharacterTrajectories,Mantis,performance,0.94,,5,71,3
+accuracy,Cricket,Mantis,performance,1.0,,5,72,3
+accuracy,DuckDuckGeese,Mantis,performance,0.407,,5,73,3
+accuracy,ERing,Mantis,performance,0.941,,5,74,3
+accuracy,EigenWorms,Mantis,performance,0.753,,5,75,3
+accuracy,Epilepsy,Mantis,performance,0.995,,5,76,3
+accuracy,EthanolConcentration,Mantis,performance,0.27,,5,77,3
+accuracy,FaceDetection,Mantis,performance,0.526,,5,78,3
+accuracy,FingerMovements,Mantis,performance,0.54,,5,79,3
+accuracy,HandMovementDirection,Mantis,performance,0.212,,5,80,3
diff --git a/result/per_paper/2502.15637/components_architecture.csv b/result/per_paper/2502.15637/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..817f1063c291680782772d6ce3b4bcc9ba99abb7
--- /dev/null
+++ b/result/per_paper/2502.15637/components_architecture.csv
@@ -0,0 +1,9 @@
+component,what_it_is,provenance,citation,evidence
+TokenGeneratorUnit,Component responsible for generating time series tokens through convolutional operations and pooling,proposed_here,,"The TokenGeneratorUnit includes conv layers, mean pooling, and instance normalization for token generation"
+Projector,Projection layer for mapping token embeddings to latent space during pre-training,proposed_here,,The projector is explicitly mentioned as part of the pre-training output pipeline
+ViTUnit,Vision Transformer architecture adapted for time series data,reused_cited,"Dosovitskiy et al., 2021","The ViTUnit contains multi-head attention, layer normalization, and positional encoding similar to original ViT"
+PredictionHead,Classifier head for downstream tasks after pre-training,proposed_here,,The prediction head is connected to the output for pre-training in the architecture diagram
+ClassToken,Special token added to the sequence for classification tasks,reused_cited,"Dosovitskiy et al., 2021",The class token is explicitly mentioned in the ViTUnit subgraph
+PositionalEncoding,Encoding mechanism to inject positional information into token embeddings,reused_cited,"Dosovitskiy et al., 2021",Positional encoding is explicitly listed in the ViTUnit subgraph
+MultiHeadAttention,Attention mechanism for capturing relationships between time series tokens,reused_cited,"Dosovitskiy et al., 2021",Multi-head attention is explicitly listed in the ViTUnit subgraph
+AdapterModules,Custom modules for handling multivariate time series data,proposed_here,,The paper states: 'we propose several adapters to handle the multivariate setting'
diff --git a/result/per_paper/2502.15637/computational.csv b/result/per_paper/2502.15637/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e313598d7fb30b201decf7ae48a606c9c4c19c37
--- /dev/null
+++ b/result/per_paper/2502.15637/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA Tesla V100-32GB GPU,,stated,pre-trained on 4 NVIDIA Tesla V100-32GB GPUs
+num_devices,4,,stated,pre-trained on 4 NVIDIA Tesla V100-32GB GPUs
+training_cost,,,not_reported,
+training_batch_size,2048,,stated,pre-trained with a batch size equal to 2048
+training_steps_or_epochs,100,,stated,pre-trained for 100 epochs
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2502.17812/accuracy_efficiency.csv b/result/per_paper/2502.17812/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..04b1cf0f6cc3f33d41da3bb35bd57f24203c39e5
--- /dev/null
+++ b/result/per_paper/2502.17812/accuracy_efficiency.csv
@@ -0,0 +1,19 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Llava-Next-72B,Llava-Next-72B,accuracy,-72,
+accuracy,Llava-Next-72B,Llava-Next-8B,accuracy,-72,
+accuracy,Llava-Next-8B,Llava-Next-72B,accuracy,-8,
+accuracy,Llava-Next-8B,Llava-Next-8B,accuracy,-8,
+accuracy,Qwen2-VL-72B,Llava-Next-72B,accuracy,2,
+accuracy,Qwen2-VL-72B,Llava-Next-8B,accuracy,2,
+accuracy,Qwen2-VL-7B,Llava-Next-72B,accuracy,2,
+accuracy,Qwen2-VL-7B,Llava-Next-8B,accuracy,2,
+accuracy,GPT-4o-mini,Llava-Next-72B,accuracy,-4,
+accuracy,GPT-4o-mini,Llava-Next-8B,accuracy,-4,
+accuracy,Gemini-1.5-Flash,Llava-Next-72B,accuracy,-1.5,
+accuracy,Gemini-1.5-Flash,Llava-Next-8B,accuracy,-1.5,
+accuracy,Point,Can MLLMs perform TSAD (Text),F1,5.88,
+accuracy,Point,Can MLLMs perform TSAD (Vision),F1,2.58,
+accuracy,Range,Can MLLMs perform TSAD (Text),F1,13.74,
+accuracy,Range,Can MLLMs perform TSAD (Vision),F1,27.45,
+accuracy,Variate,Can MLLMs perform TSAD (Text),F1,3.69,
+accuracy,Variate,Can MLLMs perform TSAD (Vision),F1,51.16,
diff --git a/result/per_paper/2502.17812/accuracy_efficiency_traced.csv b/result/per_paper/2502.17812/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5cac92fe9abd0ae1e4f40e3efbb6e8c8db5ffe8f
--- /dev/null
+++ b/result/per_paper/2502.17812/accuracy_efficiency_traced.csv
@@ -0,0 +1,19 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Llava-Next-72B,Llava-Next-72B,accuracy,-72,,1,1,0
+accuracy,Llava-Next-72B,Llava-Next-8B,accuracy,-72,,1,1,0
+accuracy,Llava-Next-8B,Llava-Next-72B,accuracy,-8,,1,2,0
+accuracy,Llava-Next-8B,Llava-Next-8B,accuracy,-8,,1,2,0
+accuracy,Qwen2-VL-72B,Llava-Next-72B,accuracy,2,,1,3,0
+accuracy,Qwen2-VL-72B,Llava-Next-8B,accuracy,2,,1,3,0
+accuracy,Qwen2-VL-7B,Llava-Next-72B,accuracy,2,,1,4,0
+accuracy,Qwen2-VL-7B,Llava-Next-8B,accuracy,2,,1,4,0
+accuracy,GPT-4o-mini,Llava-Next-72B,accuracy,-4,,1,5,0
+accuracy,GPT-4o-mini,Llava-Next-8B,accuracy,-4,,1,5,0
+accuracy,Gemini-1.5-Flash,Llava-Next-72B,accuracy,-1.5,,1,6,0
+accuracy,Gemini-1.5-Flash,Llava-Next-8B,accuracy,-1.5,,1,6,0
+accuracy,Point,Can MLLMs perform TSAD (Text),F1,5.88,,2,2,2
+accuracy,Point,Can MLLMs perform TSAD (Vision),F1,2.58,,2,2,3
+accuracy,Range,Can MLLMs perform TSAD (Text),F1,13.74,,2,3,2
+accuracy,Range,Can MLLMs perform TSAD (Vision),F1,27.45,,2,3,3
+accuracy,Variate,Can MLLMs perform TSAD (Text),F1,3.69,,2,4,2
+accuracy,Variate,Can MLLMs perform TSAD (Vision),F1,51.16,,2,4,3
diff --git a/result/per_paper/2502.17812/components_architecture.csv b/result/per_paper/2502.17812/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4289a2c7bbfd0cc3870ffd08ac34f7bb052a6da7
--- /dev/null
+++ b/result/per_paper/2502.17812/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+VisualTimeAnomaly Benchmark,"A configurable benchmark for evaluating zero-shot capabilities of MLLMs in TSAD, covering point-, range-, variate-wise, and irregular anomalies.",proposed_here,,We introduce the VisualTimeAnomaly benchmark to systematically evaluate zero-shot capability of MLLMs in TSAD... It supports flexible generation of diverse multimodal anomaly detection datasets.
+TSAD-Agents Framework,"A multi-agent system leveraging MLLMs' reasoning, planning, tool-use, and self-reflection capabilities for automatic TSAD.",proposed_here,,"We propose the first multi-agent systems TSAD-Agents for multimodal TSAD... Guided by the insights, TSAD-Agents harnesses MLLMs' strong reasoning, planning, tool-use, and self-reflection capabilities."
+Scanning Agent,An agent responsible for initial anomaly type identification and input modality selection (text/image).,proposed_here,,"The framework comprises four collaborative agents: Scanning, Detection, Planning, and Checking agents... adaptively switching between textual and visual input modalities."
+Detection Agent,An agent that invokes tools (MLLMs/traditional methods) for precise anomaly detection based on the scanning agent's input.,proposed_here,,"They synergistically coordinate to automatically reason the anomaly type, choose appropriate tools for precise anomaly detection..."
+Planning Agent,"An agent that orchestrates the detection process, including tool selection and strategy adaptation.",proposed_here,,"The framework comprises four collaborative agents... Planning agents that synergistically collaborate to reason, plan, and self-reflect."
+Checking Agent,An agent that validates detection results and refines predictions through self-reflection.,proposed_here,,"They synergistically coordinate... refine their predictions, while adaptively switching between textual and visual input modalities."
diff --git a/result/per_paper/2502.17812/computational.csv b/result/per_paper/2502.17812/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a9bde73720d83a5e6e7e3b4bf02fe6ebc30d21b1
--- /dev/null
+++ b/result/per_paper/2502.17812/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,No hardware specifications mentioned for the model.
+num_devices,,,not_reported,No information on number of devices used for training or inference.
+training_cost,,,not_reported,"No mention of training cost (e.g., monetary or computational resources)."
+training_batch_size,,,not_reported,No details on training batch size provided.
+training_steps_or_epochs,,,not_reported,No information on training duration (steps/epochs).
+precision,,,not_reported,"Precision is an evaluation metric, not a system/config parameter."
+inference_latency,,,not_reported,No latency measurements reported for inference.
+inference_throughput,,,not_reported,No throughput metrics provided for inference.
+peak_memory,,,not_reported,No memory usage details mentioned.
+flops_or_macs,,,not_reported,No computational complexity (FLOPs/MACS) reported.
+num_inference_samples,,,not_reported,No sample count for inference evaluation.
+params,,,not_reported,No parameter count for the model specified.
+context_lengths_evaluated,,,not_reported,No context length details provided.
+horizon_lengths_evaluated,,,not_reported,No horizon length evaluation metrics mentioned.
+inference_batch_size,,,not_reported,No inference batch size specified.
diff --git a/result/per_paper/2503.04118/accuracy_efficiency.csv b/result/per_paper/2503.04118/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ba9c002dc583be04762798a6887f8c62d0800d6a
--- /dev/null
+++ b/result/per_paper/2503.04118/accuracy_efficiency.csv
@@ -0,0 +1,13 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,TimeFound,MAE,0.384,
+accuracy,ETTh1,TimeFound,MAE,0.707,
+accuracy,average,TimeFound,MAE,0.768,
+accuracy,ETTh2,TimeFound,MAE,0.329,
+accuracy,ETTh2,TimeFound,MAE,0.511,
+accuracy,average,TimeFound,MAE,0.622,
+accuracy,ETTm1,TimeFound,MAE,0.320,
+accuracy,ETTm1,TimeFound,MAE,0.619,
+accuracy,average,TimeFound,MAE,0.770,
+accuracy,ETTm2,TimeFound,MAE,0.246,
+accuracy,ETTm2,TimeFound,MAE,0.415,
+accuracy,average,TimeFound,MAE,0.529,
diff --git a/result/per_paper/2503.04118/accuracy_efficiency_traced.csv b/result/per_paper/2503.04118/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b806f1ef44f97c0d18bbc1b20667c0b795acea39
--- /dev/null
+++ b/result/per_paper/2503.04118/accuracy_efficiency_traced.csv
@@ -0,0 +1,13 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,TimeFound,MAE,0.384,,1,2,5
+accuracy,ETTh1,TimeFound,MAE,0.707,,1,2,9
+accuracy,average,TimeFound,MAE,0.768,,1,6,5
+accuracy,ETTh2,TimeFound,MAE,0.329,,1,7,5
+accuracy,ETTh2,TimeFound,MAE,0.511,,1,7,9
+accuracy,average,TimeFound,MAE,0.622,,1,11,5
+accuracy,ETTm1,TimeFound,MAE,0.320,,1,12,5
+accuracy,ETTm1,TimeFound,MAE,0.619,,1,12,9
+accuracy,average,TimeFound,MAE,0.770,,1,16,5
+accuracy,ETTm2,TimeFound,MAE,0.246,,1,17,5
+accuracy,ETTm2,TimeFound,MAE,0.415,,1,17,9
+accuracy,average,TimeFound,MAE,0.529,,1,21,5
diff --git a/result/per_paper/2503.04118/components_architecture.csv b/result/per_paper/2503.04118/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..0a893f0d01776c8524d7002cd1e4bec0cae213f8
--- /dev/null
+++ b/result/per_paper/2503.04118/components_architecture.csv
@@ -0,0 +1,13 @@
+component,what_it_is,provenance,citation,evidence
+Input Module,Pre-processes raw time-series data from different domains using multi-resolution patching to capture temporal dependencies at multiple scales.,proposed_here,,"Our model begins with an Input Module that pre-processes the raw time-series from different domains, where we propose a multi-resolution patching method capable in capturing the temporal dependencies at multiple scales."
+Encoder-Decoder Transformer Blocks,Core architecture for contextual understanding of historical trends (encoder) and auto-regressive forecasting (decoder).,reused_cited,"Vaswani et al., 2017","We employ an encoder-decoder design for time series modeling and forecasting, where the encoder enables contextual understanding of historical trends while the decoder maintains the causal future prediction."
+Multi-resolution Patching,"Tokenization method that divides time series into patches of varying sizes (P₁, P₂) to capture patterns at multiple temporal scales.",proposed_here,,"We adopt a multi-resolution patching method that performs multiple divisions with different patch sizes, rather than fix-size patching."
+Projection,Component that projects multi-resolution patches into a unified representation space for processing by the encoder-decoder.,proposed_here,,"Projection --> Z1['Proj1'], Projection --> Z2['Proj2']"
+OutputHead,Generates future patch forecasts using the backbone model's outputs.,proposed_here,,OutputHead --> Backbone['Backbone']
+Backbone,Core transformer architecture that processes encoded representations and generates outputs for forecasting.,reused_cited,"Vaswani et al., 2017",Backbone --> |'o_j'| EncoderBlock
+Patch Embedding Fusion,Mechanism to combine multi-resolution patch embeddings into a unified representation for the encoder.,proposed_here,,InputModule['Input Module'] --> |'Patch Embedding Fusion'| EncoderBlock['Encoder-Decoder Transformer Blocks']
+Concatenation,"Operation that merges normalized time series with additional features (e.g., masks) for patch embedding.",proposed_here,,ValidInput --> Concatenation['Concatenation']
+Point-level Mask,Masking mechanism applied to individual time points during patch embedding to handle irregularities.,proposed_here,,Linear2 --> PointLevelMask['Point-level Mask']
+Quantile Forecasts,Output module that generates probabilistic forecasts using quantile regression.,proposed_here,,A['Point Forecasts'] --> B['Quantile Forecasts']
+SIU (Self-Inductive Unit),Component that combines linear projections of last patch representations for final forecasting.,proposed_here,,D['SIU'] --> E['Last Patch Representation']
+Normalization,"Standardization of input time series to handle varying amplitude, frequency, and stationarity across domains.",reused_cited,Z-score normalization,We apply the commonly used standard sc
diff --git a/result/per_paper/2503.04118/computational.csv b/result/per_paper/2503.04118/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..41fbb17fc4b9a345af11bf8a1f84fd8a47ae6cb3
--- /dev/null
+++ b/result/per_paper/2503.04118/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,1024,samples,stated,"trained for 200K steps, with a batch size of 1024"
+training_steps_or_epochs,200000,steps,stated,trained for 200K steps
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,512,points,stated,context length is fixed at 512
+horizon_lengths_evaluated,"[96, 192, 336, 720]",points,stated,"horizon lengths of {96, 192, 336 and 720}"
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2505.13033/accuracy_efficiency.csv b/result/per_paper/2505.13033/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6ed55c1ec4b0d3e9fdb2ac49488b75b9c1772474
--- /dev/null
+++ b/result/per_paper/2505.13033/accuracy_efficiency.csv
@@ -0,0 +1,111 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,TSPulse,MR,0.146,
+accuracy,ETTh1,TSPulse,MR,0.042,
+accuracy,ETTh2,TSPulse,MR,0.065,
+accuracy,ETTh2,TSPulse,MR,0.041,
+accuracy,ETTm1,TSPulse,MR,0.049,
+accuracy,ETTm1,TSPulse,MR,0.018,
+accuracy,ETTm2,TSPulse,MR,0.034,
+accuracy,ETTm2,TSPulse,MR,0.020,
+accuracy,Weather,TSPulse,MR,0.036,
+accuracy,Weather,TSPulse,MR,0.036,
+accuracy,Electricity,TSPulse,MR,0.070,
+accuracy,Electricity,TSPulse,MR,0.043,
+accuracy,ETTh1,TSPulse,MR,0.209,
+accuracy,ETTh1,TSPulse,MR,0.055,
+accuracy,ETTh2,TSPulse,MR,0.081,
+accuracy,ETTh2,TSPulse,MR,0.050,
+accuracy,ETTm1,TSPulse,MR,0.068,
+accuracy,ETTm1,TSPulse,MR,0.023,
+accuracy,ETTm2,TSPulse,MR,0.046,
+accuracy,ETTm2,TSPulse,MR,0.022,
+accuracy,Weather,TSPulse,MR,0.048,
+accuracy,Weather,TSPulse,MR,0.034,
+accuracy,Electricity,TSPulse,MR,0.084,
+accuracy,Electricity,TSPulse,MR,0.053,
+accuracy,ArticularyWordRecognition,TSPulse,Accuracy,0.98,
+accuracy,AtrialFibrillation,TSPulse,Accuracy,0.467,
+accuracy,BasicMotions,TSPulse,Accuracy,1.0,
+accuracy,CharacterTrajectories,TSPulse,Accuracy,0.987,
+accuracy,Cricket,TSPulse,Accuracy,0.917,
+accuracy,DuckDuckGeese,TSPulse,Accuracy,0.72,
+accuracy,ERing,TSPulse,Accuracy,0.937,
+accuracy,Epilepsy,TSPulse,Accuracy,0.971,
+accuracy,EthanolConcentration,TSPulse,Accuracy,0.247,
+accuracy,FaceDetection,TSPulse,Accuracy,0.675,
+accuracy,FingerMovements,TSPulse,Accuracy,0.53,
+accuracy,HandMovementDirection,TSPulse,Accuracy,0.649,
+accuracy,Handwriting,TSPulse,Accuracy,0.215,
+accuracy,Heartbeat,TSPulse,Accuracy,0.702,
+accuracy,InsectWingbeat,TSPulse,Accuracy,0.723,
+accuracy,JapaneseVowels,TSPulse,Accuracy,0.976,
+accuracy,LSST,TSPulse,Accuracy,0.518,
+accuracy,Libras,TSPulse,Accuracy,0.767,
+accuracy,MotorImagery,TSPulse,Accuracy,0.58,
+accuracy,NATOPS,TSPulse,Accuracy,0.878,
+accuracy,PEMS-SF,TSPulse,Accuracy,0.855,
+accuracy,PenDigits,TSPulse,Accuracy,0.969,
+accuracy,PhonemeSpectra,TSPulse,Accuracy,0.156,
+accuracy,RacketSports,TSPulse,Accuracy,0.901,
+accuracy,SelfRegulationSCP1,TSPulse,Accuracy,0.836,
+accuracy,SelfRegulationSCP2,TSPulse,Accuracy,0.511,
+accuracy,SpokenArabicDigits,TSPulse,Accuracy,0.984,
+accuracy,StandWalkJump,TSPulse,Accuracy,0.733,
+accuracy,UWaveGestureLibrary,TSPulse,Accuracy,0.881,
+accuracy,Mean,TSPulse,Accuracy,0.733,
+accuracy,IMP(%),TSPulse,Accuracy,5,
+accuracy,ETTh1,TSPulse w/o Dual_space,MAE,0.227,
+accuracy,ETTh1,TSPulse,MAE,0.209,
+accuracy,ETTh1,TSPulse w/o Hybrid Pre-training,MAE,0.189,
+accuracy,ETTh1,TSPulse w/o Dual_space,MAE,0.158,
+accuracy,ETTh1,TSPulse,MAE,0.146,
+accuracy,ETTh1,TSPulse w/o Hybrid Pre-training,MAE,0.498,
+accuracy,ETTh2,TSPulse w/o Dual_space,MAE,0.086,
+accuracy,ETTh2,TSPulse,MAE,0.081,
+accuracy,ETTh2,TSPulse w/o Hybrid Pre-training,MAE,0.077,
+accuracy,ETTh2,TSPulse w/o Dual_space,MAE,0.069,
+accuracy,ETTh2,TSPulse,MAE,0.065,
+accuracy,ETTh2,TSPulse w/o Hybrid Pre-training,MAE,0.216,
+accuracy,ETTm1,TSPulse w/o Dual_space,MAE,0.078,
+accuracy,ETTm1,TSPulse,MAE,0.068,
+accuracy,ETTm1,TSPulse w/o Hybrid Pre-training,MAE,0.065,
+accuracy,ETTm1,TSPulse w/o Dual_space,MAE,0.055,
+accuracy,ETTm1,TSPulse,MAE,0.049,
+accuracy,ETTm1,TSPulse w/o Hybrid Pre-training,MAE,0.449,
+accuracy,ETTm2,TSPulse w/o Dual_space,MAE,0.048,
+accuracy,ETTm2,TSPulse,MAE,0.046,
+accuracy,ETTm2,TSPulse w/o Hybrid Pre-training,MAE,0.044,
+accuracy,ETTm2,TSPulse w/o Dual_space,MAE,0.036,
+accuracy,ETTm2,TSPulse,MAE,0.034,
+accuracy,ETTm2,TSPulse w/o Hybrid Pre-training,MAE,0.154,
+accuracy,Weather,TSPulse w/o Dual_space,MAE,0.051,
+accuracy,Weather,TSPulse,MAE,0.048,
+accuracy,Weather,TSPulse w/o Hybrid Pre-training,MAE,0.048,
+accuracy,Weather,TSPulse w/o Dual_space,MAE,0.038,
+accuracy,Weather,TSPulse,MAE,0.036,
+accuracy,Weather,TSPulse w/o Hybrid Pre-training,MAE,0.146,
+accuracy,Electricity,TSPulse w/o Dual_space,MAE,0.105,
+accuracy,Electricity,TSPulse,MAE,0.084,
+accuracy,Electricity,TSPulse w/o Hybrid Pre-training,MAE,0.069,
+accuracy,Electricity,TSPulse w/o Dual_space,MAE,0.085,
+accuracy,Electricity,TSPulse,MAE,0.070,
+accuracy,Electricity,TSPulse w/o Hybrid Pre-training,MAE,0.488,
+accuracy,ArticularyWordRecognition,TSPulse,accuracy,0.98,
+accuracy,AtrialFibrillation,TSPulse,accuracy,0.467,
+accuracy,BasicMotions,TSPulse,accuracy,1.0,
+accuracy,Cricket,TSPulse,accuracy,0.917,
+accuracy,ERing,TSPulse,accuracy,0.937,
+accuracy,Epilepsy,TSPulse,accuracy,0.971,
+accuracy,EthanolConcentration,TSPulse,accuracy,0.247,
+accuracy,FingerMovements,TSPulse,accuracy,0.53,
+accuracy,HandMovementDirection,TSPulse,accuracy,0.649,
+accuracy,Handwriting,TSPulse,accuracy,0.215,
+accuracy,JapaneseVowels,TSPulse,accuracy,0.976,
+accuracy,MotorImagery,TSPulse,accuracy,0.58,
+accuracy,NATOPS,TSPulse,accuracy,0.878,
+accuracy,RacketSports,TSPulse,accuracy,0.901,
+accuracy,SelfRegulationSCP1,TSPulse,accuracy,0.836,
+accuracy,StandWalkJump,TSPulse,accuracy,0.733,
+accuracy,UWaveGestureLibrary,TSPulse,accuracy,0.881,
+accuracy,Mean,TSPulse,accuracy,0.747,
+accuracy,IMP(%),TSPulse,accuracy,8,
diff --git a/result/per_paper/2505.13033/accuracy_efficiency_traced.csv b/result/per_paper/2505.13033/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..8ff7377c2215ba93603056a384cd2ede880887b3
--- /dev/null
+++ b/result/per_paper/2505.13033/accuracy_efficiency_traced.csv
@@ -0,0 +1,111 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,TSPulse,MR,0.146,,2,2,2
+accuracy,ETTh1,TSPulse,MR,0.042,,2,2,9
+accuracy,ETTh2,TSPulse,MR,0.065,,2,7,2
+accuracy,ETTh2,TSPulse,MR,0.041,,2,7,9
+accuracy,ETTm1,TSPulse,MR,0.049,,2,12,2
+accuracy,ETTm1,TSPulse,MR,0.018,,2,12,9
+accuracy,ETTm2,TSPulse,MR,0.034,,2,17,2
+accuracy,ETTm2,TSPulse,MR,0.020,,2,17,9
+accuracy,Weather,TSPulse,MR,0.036,,2,22,2
+accuracy,Weather,TSPulse,MR,0.036,,2,22,9
+accuracy,Electricity,TSPulse,MR,0.070,,2,27,2
+accuracy,Electricity,TSPulse,MR,0.043,,2,27,9
+accuracy,ETTh1,TSPulse,MR,0.209,,3,2,2
+accuracy,ETTh1,TSPulse,MR,0.055,,3,2,9
+accuracy,ETTh2,TSPulse,MR,0.081,,3,7,2
+accuracy,ETTh2,TSPulse,MR,0.050,,3,7,9
+accuracy,ETTm1,TSPulse,MR,0.068,,3,12,2
+accuracy,ETTm1,TSPulse,MR,0.023,,3,12,9
+accuracy,ETTm2,TSPulse,MR,0.046,,3,17,2
+accuracy,ETTm2,TSPulse,MR,0.022,,3,17,9
+accuracy,Weather,TSPulse,MR,0.048,,3,22,2
+accuracy,Weather,TSPulse,MR,0.034,,3,22,9
+accuracy,Electricity,TSPulse,MR,0.084,,3,27,2
+accuracy,Electricity,TSPulse,MR,0.053,,3,27,9
+accuracy,ArticularyWordRecognition,TSPulse,Accuracy,0.98,,4,1,1
+accuracy,AtrialFibrillation,TSPulse,Accuracy,0.467,,4,2,1
+accuracy,BasicMotions,TSPulse,Accuracy,1.0,,4,3,1
+accuracy,CharacterTrajectories,TSPulse,Accuracy,0.987,,4,4,1
+accuracy,Cricket,TSPulse,Accuracy,0.917,,4,5,1
+accuracy,DuckDuckGeese,TSPulse,Accuracy,0.72,,4,6,1
+accuracy,ERing,TSPulse,Accuracy,0.937,,4,7,1
+accuracy,Epilepsy,TSPulse,Accuracy,0.971,,4,8,1
+accuracy,EthanolConcentration,TSPulse,Accuracy,0.247,,4,9,1
+accuracy,FaceDetection,TSPulse,Accuracy,0.675,,4,10,1
+accuracy,FingerMovements,TSPulse,Accuracy,0.53,,4,11,1
+accuracy,HandMovementDirection,TSPulse,Accuracy,0.649,,4,12,1
+accuracy,Handwriting,TSPulse,Accuracy,0.215,,4,13,1
+accuracy,Heartbeat,TSPulse,Accuracy,0.702,,4,14,1
+accuracy,InsectWingbeat,TSPulse,Accuracy,0.723,,4,15,1
+accuracy,JapaneseVowels,TSPulse,Accuracy,0.976,,4,16,1
+accuracy,LSST,TSPulse,Accuracy,0.518,,4,17,1
+accuracy,Libras,TSPulse,Accuracy,0.767,,4,18,1
+accuracy,MotorImagery,TSPulse,Accuracy,0.58,,4,19,1
+accuracy,NATOPS,TSPulse,Accuracy,0.878,,4,20,1
+accuracy,PEMS-SF,TSPulse,Accuracy,0.855,,4,21,1
+accuracy,PenDigits,TSPulse,Accuracy,0.969,,4,22,1
+accuracy,PhonemeSpectra,TSPulse,Accuracy,0.156,,4,23,1
+accuracy,RacketSports,TSPulse,Accuracy,0.901,,4,24,1
+accuracy,SelfRegulationSCP1,TSPulse,Accuracy,0.836,,4,25,1
+accuracy,SelfRegulationSCP2,TSPulse,Accuracy,0.511,,4,26,1
+accuracy,SpokenArabicDigits,TSPulse,Accuracy,0.984,,4,27,1
+accuracy,StandWalkJump,TSPulse,Accuracy,0.733,,4,28,1
+accuracy,UWaveGestureLibrary,TSPulse,Accuracy,0.881,,4,29,1
+accuracy,Mean,TSPulse,Accuracy,0.733,,4,30,1
+accuracy,IMP(%),TSPulse,Accuracy,5,,4,31,1
+accuracy,ETTh1,TSPulse w/o Dual_space,MAE,0.227,,5,2,2
+accuracy,ETTh1,TSPulse,MAE,0.209,,5,2,3
+accuracy,ETTh1,TSPulse w/o Hybrid Pre-training,MAE,0.189,,5,2,4
+accuracy,ETTh1,TSPulse w/o Dual_space,MAE,0.158,,5,2,5
+accuracy,ETTh1,TSPulse,MAE,0.146,,5,2,6
+accuracy,ETTh1,TSPulse w/o Hybrid Pre-training,MAE,0.498,,5,2,7
+accuracy,ETTh2,TSPulse w/o Dual_space,MAE,0.086,,5,7,2
+accuracy,ETTh2,TSPulse,MAE,0.081,,5,7,3
+accuracy,ETTh2,TSPulse w/o Hybrid Pre-training,MAE,0.077,,5,7,4
+accuracy,ETTh2,TSPulse w/o Dual_space,MAE,0.069,,5,7,5
+accuracy,ETTh2,TSPulse,MAE,0.065,,5,7,6
+accuracy,ETTh2,TSPulse w/o Hybrid Pre-training,MAE,0.216,,5,7,7
+accuracy,ETTm1,TSPulse w/o Dual_space,MAE,0.078,,5,12,2
+accuracy,ETTm1,TSPulse,MAE,0.068,,5,12,3
+accuracy,ETTm1,TSPulse w/o Hybrid Pre-training,MAE,0.065,,5,12,4
+accuracy,ETTm1,TSPulse w/o Dual_space,MAE,0.055,,5,12,5
+accuracy,ETTm1,TSPulse,MAE,0.049,,5,12,6
+accuracy,ETTm1,TSPulse w/o Hybrid Pre-training,MAE,0.449,,5,12,7
+accuracy,ETTm2,TSPulse w/o Dual_space,MAE,0.048,,5,17,2
+accuracy,ETTm2,TSPulse,MAE,0.046,,5,17,3
+accuracy,ETTm2,TSPulse w/o Hybrid Pre-training,MAE,0.044,,5,17,4
+accuracy,ETTm2,TSPulse w/o Dual_space,MAE,0.036,,5,17,5
+accuracy,ETTm2,TSPulse,MAE,0.034,,5,17,6
+accuracy,ETTm2,TSPulse w/o Hybrid Pre-training,MAE,0.154,,5,17,7
+accuracy,Weather,TSPulse w/o Dual_space,MAE,0.051,,5,22,2
+accuracy,Weather,TSPulse,MAE,0.048,,5,22,3
+accuracy,Weather,TSPulse w/o Hybrid Pre-training,MAE,0.048,,5,22,4
+accuracy,Weather,TSPulse w/o Dual_space,MAE,0.038,,5,22,5
+accuracy,Weather,TSPulse,MAE,0.036,,5,22,6
+accuracy,Weather,TSPulse w/o Hybrid Pre-training,MAE,0.146,,5,22,7
+accuracy,Electricity,TSPulse w/o Dual_space,MAE,0.105,,5,27,2
+accuracy,Electricity,TSPulse,MAE,0.084,,5,27,3
+accuracy,Electricity,TSPulse w/o Hybrid Pre-training,MAE,0.069,,5,27,4
+accuracy,Electricity,TSPulse w/o Dual_space,MAE,0.085,,5,27,5
+accuracy,Electricity,TSPulse,MAE,0.070,,5,27,6
+accuracy,Electricity,TSPulse w/o Hybrid Pre-training,MAE,0.488,,5,27,7
+accuracy,ArticularyWordRecognition,TSPulse,accuracy,0.98,,6,1,1
+accuracy,AtrialFibrillation,TSPulse,accuracy,0.467,,6,2,1
+accuracy,BasicMotions,TSPulse,accuracy,1.0,,6,3,1
+accuracy,Cricket,TSPulse,accuracy,0.917,,6,4,1
+accuracy,ERing,TSPulse,accuracy,0.937,,6,5,1
+accuracy,Epilepsy,TSPulse,accuracy,0.971,,6,6,1
+accuracy,EthanolConcentration,TSPulse,accuracy,0.247,,6,7,1
+accuracy,FingerMovements,TSPulse,accuracy,0.53,,6,8,1
+accuracy,HandMovementDirection,TSPulse,accuracy,0.649,,6,9,1
+accuracy,Handwriting,TSPulse,accuracy,0.215,,6,10,1
+accuracy,JapaneseVowels,TSPulse,accuracy,0.976,,6,11,1
+accuracy,MotorImagery,TSPulse,accuracy,0.58,,6,12,1
+accuracy,NATOPS,TSPulse,accuracy,0.878,,6,13,1
+accuracy,RacketSports,TSPulse,accuracy,0.901,,6,14,1
+accuracy,SelfRegulationSCP1,TSPulse,accuracy,0.836,,6,15,1
+accuracy,StandWalkJump,TSPulse,accuracy,0.733,,6,16,1
+accuracy,UWaveGestureLibrary,TSPulse,accuracy,0.881,,6,17,1
+accuracy,Mean,TSPulse,accuracy,0.747,,6,18,1
+accuracy,IMP(%),TSPulse,accuracy,8,,6,19,1
diff --git a/result/per_paper/2505.13033/components_architecture.csv b/result/per_paper/2505.13033/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2fbf43a5e6b0b9ac74dca7c2d877c5633732af56
--- /dev/null
+++ b/result/per_paper/2505.13033/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Pre-training Framework with Disentanglement,"A novel pre-training framework that augments masked reconstruction with explicit disentanglement across representation spaces and abstraction levels, enabling the learning of three complementary embedding views (temporal, spectral, and semantic).",proposed_here,,"TSPulse introduces a novel pre-training framework that augments masked reconstruction with explicit disentanglement across spaces and abstractions, learning three complementary embedding views (temporal, spectral, and semantic) to effectively enable zero-shot transfer."
+Temporal Embeddings,"Detailed embeddings for fine-grained time analysis, capturing local patterns and irregularities in the time domain.",proposed_here,,explicitly producing three distinct types of embeddings during pre-training: (i) detailed temporal embeddings for fine-grained time analysis...
+Spectral Embeddings,"Detailed embeddings for frequency-aware fidelity, capturing periodic patterns and structures in the frequency domain.",proposed_here,,explicitly producing three distinct types of embeddings during pre-training: (ii) detailed spectral embeddings for frequency-aware fidelity...
+Semantic Embeddings,"Embeddings for high-level task understanding, capturing global semantics and abstract representations of time-series data.",proposed_here,,explicitly producing three distinct types of embeddings during pre-training: (iii) semantic embeddings for high-level task understanding.
+Post-hoc Fusers,"Lightweight modules that selectively attend and fuse the disentangled views (temporal, spectral, semantic) based on task type, enabling task-specific specialization.",proposed_here,,"we introduce various lightweight post-hoc fusers that selectively attend and fuse these disentangled views based on task type, enabling simple but effective task specializations."
+Hybrid Masking Strategy,"A simple yet effective masking strategy that enhances missing diversity during pre-training, improving robustness and mitigating mask-induced bias.",proposed_here,,we propose a simple yet effective hybrid masking strategy that enhances missing diversity during pre-training.
diff --git a/result/per_paper/2505.13033/computational.csv b/result/per_paper/2505.13033/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f66286f2e566495c89b2c6f30da162b1c480fbe6
--- /dev/null
+++ b/result/per_paper/2505.13033/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,A100,GPU,stated,Pre-training on 1B samples takes just one day with 8×A100 GPUs
+num_devices,8,units,stated,Pre-training on 1B samples takes just one day with 8×A100 GPUs
+training_cost,1,day,stated,Pre-training on 1B samples takes just one day with 8×A100 GPUs
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,7.16,ms,stated,GPU Inference Time (ms): 7.16
+inference_throughput,,,not_reported,
+peak_memory,0.39,GB,stated,Max. Memory (GB): 0.39
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,1.06,M,stated,Params (M): 1.06
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2505.14766/accuracy_efficiency.csv b/result/per_paper/2505.14766/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..353b00edf4976b530155ad55d98a509922bba757
--- /dev/null
+++ b/result/per_paper/2505.14766/accuracy_efficiency.csv
@@ -0,0 +1,157 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ETTh1,BOOM,MAE,0.381,
+accuracy,MSE ↓,BOOM,MAE,0.384,
+accuracy,MSE ↓,BOOM,MAE,0.425,
+accuracy,MSE ↓,BOOM,MAE,0.456,
+accuracy,MSE ↓,BOOM,MAE,0.470,
+accuracy,ETTh2,BOOM,MAE,0.310,
+accuracy,MSE ↓,BOOM,MAE,0.277,
+accuracy,MSE ↓,BOOM,MAE,0.340,
+accuracy,MSE ↓,BOOM,MAE,0.371,
+accuracy,MSE ↓,BOOM,MAE,0.394,
+accuracy,ETTm1,BOOM,MAE,0.333,
+accuracy,MSE ↓,BOOM,MAE,0.335,
+accuracy,MSE ↓,BOOM,MAE,0.366,
+accuracy,MSE ↓,BOOM,MAE,0.391,
+accuracy,MSE ↓,BOOM,MAE,0.434,
+accuracy,ETTm2,BOOM,MAE,0.237,
+accuracy,MSE ↓,BOOM,MAE,0.195,
+accuracy,MSE ↓,BOOM,MAE,0.247,
+accuracy,MSE ↓,BOOM,MAE,0.291,
+accuracy,MSE ↓,BOOM,MAE,0.355,
+accuracy,Electricity,BOOM,MAE,0.213,
+accuracy,MSE ↓,BOOM,MAE,0.158,
+accuracy,MSE ↓,BOOM,MAE,0.174,
+accuracy,MSE ↓,BOOM,MAE,0.191,
+accuracy,MSE ↓,BOOM,MAE,0.229,
+accuracy,Weather,BOOM,MAE,0.179,
+accuracy,MSE ↓,BOOM,MAE,0.167,
+accuracy,MSE ↓,BOOM,MAE,0.209,
+accuracy,MSE ↓,BOOM,MAE,0.256,
+accuracy,MSE ↓,BOOM,MAE,0.321,
+accuracy,Best Count,BOOM,MAE,29,
+accuracy,Application usage,BOOM,MASE and CRPS,0.639,
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.440,
+accuracy,Database,BOOM,MASE and CRPS,0.635,
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.429,
+accuracy,Infrastructure,BOOM,MASE and CRPS,0.568,
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.476,
+accuracy,Networking,BOOM,MASE and CRPS,0.635,
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.493,
+accuracy,Security,BOOM,MASE and CRPS,0.682,
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.505,
+accuracy,ETTh1,Toto,MAE,0.413,
+accuracy,ETTh1,Moirai_Small,MAE,0.424,
+accuracy,ETTh1,Moirai_Base,MAE,0.438,
+accuracy,ETTh1,Moirai_Large,MAE,0.469,
+accuracy,ETTh1,VisionTS,MAE,0.414,
+accuracy,ETTh1,Time-MoE_Base,MAE,0.424,
+accuracy,ETTh1,Time-MoE_Large,MAE,0.419,
+accuracy,ETTh1,Time-MoE_Ultra,MAE,0.426,
+accuracy,MSE ↓,Toto,MAE,0.400,
+accuracy,MSE ↓,Moirai_Small,MAE,0.434,
+accuracy,MSE ↓,Moirai_Base,MAE,0.510,
+accuracy,MSE ↓,Moirai_Large,MAE,0.390,
+accuracy,MSE ↓,VisionTS,MAE,0.400,
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.394,
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.412,
+accuracy,ETTh2,Toto,MAE,0.363,
+accuracy,ETTh2,Moirai_Small,MAE,0.379,
+accuracy,ETTh2,Moirai_Base,MAE,0.382,
+accuracy,ETTh2,Moirai_Large,MAE,0.376,
+accuracy,ETTh2,VisionTS,MAE,0.375,
+accuracy,ETTh2,Time-MoE_Base,MAE,0.404,
+accuracy,ETTh2,Time-MoE_Large,MAE,0.415,
+accuracy,ETTh2,Time-MoE_Ultra,MAE,0.399,
+accuracy,MSE ↓,Toto,MAE,0.341,
+accuracy,MSE ↓,Moirai_Small,MAE,0.345,
+accuracy,MSE ↓,Moirai_Base,MAE,0.354,
+accuracy,MSE ↓,Moirai_Large,MAE,0.333,
+accuracy,MSE ↓,VisionTS,MAE,0.366,
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.405,
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.371,
+accuracy,ETTm1,Toto,MAE,0.378,
+accuracy,ETTm1,Moirai_Small,MAE,0.409,
+accuracy,ETTm1,Moirai_Base,MAE,0.388,
+accuracy,ETTm1,Moirai_Large,MAE,0.389,
+accuracy,ETTm1,VisionTS,MAE,0.372,
+accuracy,ETTm1,Time-MoE_Base,MAE,0.415,
+accuracy,ETTm1,Time-MoE_Large,MAE,0.405,
+accuracy,ETTm1,Time-MoE_Ultra,MAE,0.391,
+accuracy,MSE ↓,Toto,MAE,0.448,
+accuracy,MSE ↓,Moirai_Small,MAE,0.381,
+accuracy,MSE ↓,Moirai_Base,MAE,0.390,
+accuracy,MSE ↓,Moirai_Large,MAE,0.374,
+accuracy,MSE ↓,VisionTS,MAE,0.394,
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.376,
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.356,
+accuracy,ETTm2,Toto,MAE,0.303,
+accuracy,ETTm2,Moirai_Small,MAE,0.341,
+accuracy,ETTm2,Moirai_Base,MAE,0.321,
+accuracy,ETTm2,Moirai_Large,MAE,0.320,
+accuracy,ETTm2,VisionTS,MAE,0.321,
+accuracy,ETTm2,Time-MoE_Base,MAE,0.365,
+accuracy,ETTm2,Time-MoE_Large,MAE,0.361,
+accuracy,ETTm2,Time-MoE_Ultra,MAE,0.344,
+accuracy,MSE ↓,Toto,MAE,0.300,
+accuracy,MSE ↓,Moirai_Small,MAE,0.272,
+accuracy,MSE ↓,Moirai_Base,MAE,0.276,
+accuracy,MSE ↓,Moirai_Large,MAE,0.282,
+accuracy,MSE ↓,VisionTS,MAE,0.317,
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.316,
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.288,
+accuracy,Electricity,Toto,MAE,0.243,
+accuracy,Electricity,Moirai_Small,MAE,0.320,
+accuracy,Electricity,Moirai_Base,MAE,0.274,
+accuracy,Electricity,Moirai_Large,MAE,0.273,
+accuracy,Electricity,VisionTS,MAE,0.294,
+accuracy,MSE ↓,Toto,MAE,0.233,
+accuracy,MSE ↓,Moirai_Small,MAE,0.188,
+accuracy,MSE ↓,Moirai_Base,MAE,0.188,
+accuracy,MSE ↓,Moirai_Large,MAE,0.207,
+accuracy,Weather,Toto,MAE,0.245,
+accuracy,Weather,Moirai_Small,MAE,0.267,
+accuracy,Weather,Moirai_Base,MAE,0.261,
+accuracy,Weather,Moirai_Large,MAE,0.275,
+accuracy,Weather,VisionTS,MAE,0.292,
+accuracy,Weather,Time-MoE_Base,MAE,0.297,
+accuracy,Weather,Time-MoE_Large,MAE,0.300,
+accuracy,Weather,Time-MoE_Ultra,MAE,0.288,
+accuracy,MSE ↓,Toto,MAE,0.242,
+accuracy,MSE ↓,Moirai_Small,MAE,0.238,
+accuracy,MSE ↓,Moirai_Base,MAE,0.259,
+accuracy,MSE ↓,Moirai_Large,MAE,0.269,
+accuracy,MSE ↓,VisionTS,MAE,0.265,
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.270,
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.256,
+accuracy,Mean,Toto,MAE,0.324,
+accuracy,Mean,Moirai_Small,MAE,0.357,
+accuracy,Mean,Moirai_Base,MAE,0.344,
+accuracy,Mean,Moirai_Large,MAE,0.350,
+accuracy,Mean,VisionTS,MAE,0.345,
+accuracy,MSE ↓,Toto,MAE,0.327,
+accuracy,MSE ↓,Moirai_Small,MAE,0.310,
+accuracy,MSE ↓,Moirai_Base,MAE,0.330,
+accuracy,MSE ↓,Moirai_Large,MAE,0.309,
+accuracy,Best Count,Toto,MAE,8,
+accuracy,Best Count,Moirai_Small,MAE,0,
+accuracy,Best Count,Moirai_Base,MAE,0,
+accuracy,Best Count,Moirai_Large,MAE,0,
+accuracy,Best Count,VisionTS,MAE,3,
+accuracy,Best Count,Time-MoE_Base,MAE,0,
+accuracy,Best Count,Time-MoE_Large,MAE,0,
+accuracy,Best Count,Time-MoE_Ultra,MAE,1,
+accuracy,Count,BOOM,"MASE, CRPS",0.687,
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.370,
+accuracy,Distribution,BOOM,"MASE, CRPS",0.658,
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.434,
+accuracy,Gauge,BOOM,"MASE, CRPS",0.583,
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.471,
+accuracy,Rate,BOOM,"MASE, CRPS",0.634,
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.433,
+accuracy,BOOM,BOOM,MASE,0.617,
+accuracy,CRPS $\downarrow$,BOOM,MASE,0.442,
+accuracy,Rank $\downarrow$,BOOM,MASE,4.905,
+accuracy,BOOMLET,BOOM,MASE,0.617,
+accuracy,CRPS $\downarrow$,BOOM,MASE,0.631,
+accuracy,Rank $\downarrow$,BOOM,MASE,5.711,
diff --git a/result/per_paper/2505.14766/accuracy_efficiency_traced.csv b/result/per_paper/2505.14766/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6dc7b513bc7b3c5d5f7bc565ba7ecd4a1f06dc2a
--- /dev/null
+++ b/result/per_paper/2505.14766/accuracy_efficiency_traced.csv
@@ -0,0 +1,157 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ETTh1,BOOM,MAE,0.381,,2,2,3
+accuracy,MSE ↓,BOOM,MAE,0.384,,2,3,3
+accuracy,MSE ↓,BOOM,MAE,0.425,,2,5,3
+accuracy,MSE ↓,BOOM,MAE,0.456,,2,7,3
+accuracy,MSE ↓,BOOM,MAE,0.470,,2,9,3
+accuracy,ETTh2,BOOM,MAE,0.310,,2,10,3
+accuracy,MSE ↓,BOOM,MAE,0.277,,2,11,3
+accuracy,MSE ↓,BOOM,MAE,0.340,,2,13,3
+accuracy,MSE ↓,BOOM,MAE,0.371,,2,15,3
+accuracy,MSE ↓,BOOM,MAE,0.394,,2,17,3
+accuracy,ETTm1,BOOM,MAE,0.333,,2,18,3
+accuracy,MSE ↓,BOOM,MAE,0.335,,2,19,3
+accuracy,MSE ↓,BOOM,MAE,0.366,,2,21,3
+accuracy,MSE ↓,BOOM,MAE,0.391,,2,23,3
+accuracy,MSE ↓,BOOM,MAE,0.434,,2,25,3
+accuracy,ETTm2,BOOM,MAE,0.237,,2,26,3
+accuracy,MSE ↓,BOOM,MAE,0.195,,2,27,3
+accuracy,MSE ↓,BOOM,MAE,0.247,,2,29,3
+accuracy,MSE ↓,BOOM,MAE,0.291,,2,31,3
+accuracy,MSE ↓,BOOM,MAE,0.355,,2,33,3
+accuracy,Electricity,BOOM,MAE,0.213,,2,34,3
+accuracy,MSE ↓,BOOM,MAE,0.158,,2,35,3
+accuracy,MSE ↓,BOOM,MAE,0.174,,2,37,3
+accuracy,MSE ↓,BOOM,MAE,0.191,,2,39,3
+accuracy,MSE ↓,BOOM,MAE,0.229,,2,41,3
+accuracy,Weather,BOOM,MAE,0.179,,2,42,3
+accuracy,MSE ↓,BOOM,MAE,0.167,,2,43,3
+accuracy,MSE ↓,BOOM,MAE,0.209,,2,45,3
+accuracy,MSE ↓,BOOM,MAE,0.256,,2,47,3
+accuracy,MSE ↓,BOOM,MAE,0.321,,2,49,3
+accuracy,Best Count,BOOM,MAE,29,,2,50,3
+accuracy,Application usage,BOOM,MASE and CRPS,0.639,,4,1,2
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.440,,4,2,2
+accuracy,Database,BOOM,MASE and CRPS,0.635,,4,3,2
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.429,,4,4,2
+accuracy,Infrastructure,BOOM,MASE and CRPS,0.568,,4,5,2
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.476,,4,6,2
+accuracy,Networking,BOOM,MASE and CRPS,0.635,,4,7,2
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.493,,4,8,2
+accuracy,Security,BOOM,MASE and CRPS,0.682,,4,9,2
+accuracy,CRPS ↓,BOOM,MASE and CRPS,0.505,,4,10,2
+accuracy,ETTh1,Toto,MAE,0.413,,5,2,2
+accuracy,ETTh1,Moirai_Small,MAE,0.424,,5,2,3
+accuracy,ETTh1,Moirai_Base,MAE,0.438,,5,2,4
+accuracy,ETTh1,Moirai_Large,MAE,0.469,,5,2,5
+accuracy,ETTh1,VisionTS,MAE,0.414,,5,2,6
+accuracy,ETTh1,Time-MoE_Base,MAE,0.424,,5,2,7
+accuracy,ETTh1,Time-MoE_Large,MAE,0.419,,5,2,8
+accuracy,ETTh1,Time-MoE_Ultra,MAE,0.426,,5,2,9
+accuracy,MSE ↓,Toto,MAE,0.400,,5,3,2
+accuracy,MSE ↓,Moirai_Small,MAE,0.434,,5,3,3
+accuracy,MSE ↓,Moirai_Base,MAE,0.510,,5,3,4
+accuracy,MSE ↓,Moirai_Large,MAE,0.390,,5,3,5
+accuracy,MSE ↓,VisionTS,MAE,0.400,,5,3,6
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.394,,5,3,7
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.412,,5,3,8
+accuracy,ETTh2,Toto,MAE,0.363,,5,4,2
+accuracy,ETTh2,Moirai_Small,MAE,0.379,,5,4,3
+accuracy,ETTh2,Moirai_Base,MAE,0.382,,5,4,4
+accuracy,ETTh2,Moirai_Large,MAE,0.376,,5,4,5
+accuracy,ETTh2,VisionTS,MAE,0.375,,5,4,6
+accuracy,ETTh2,Time-MoE_Base,MAE,0.404,,5,4,7
+accuracy,ETTh2,Time-MoE_Large,MAE,0.415,,5,4,8
+accuracy,ETTh2,Time-MoE_Ultra,MAE,0.399,,5,4,9
+accuracy,MSE ↓,Toto,MAE,0.341,,5,5,2
+accuracy,MSE ↓,Moirai_Small,MAE,0.345,,5,5,3
+accuracy,MSE ↓,Moirai_Base,MAE,0.354,,5,5,4
+accuracy,MSE ↓,Moirai_Large,MAE,0.333,,5,5,5
+accuracy,MSE ↓,VisionTS,MAE,0.366,,5,5,6
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.405,,5,5,7
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.371,,5,5,8
+accuracy,ETTm1,Toto,MAE,0.378,,5,6,2
+accuracy,ETTm1,Moirai_Small,MAE,0.409,,5,6,3
+accuracy,ETTm1,Moirai_Base,MAE,0.388,,5,6,4
+accuracy,ETTm1,Moirai_Large,MAE,0.389,,5,6,5
+accuracy,ETTm1,VisionTS,MAE,0.372,,5,6,6
+accuracy,ETTm1,Time-MoE_Base,MAE,0.415,,5,6,7
+accuracy,ETTm1,Time-MoE_Large,MAE,0.405,,5,6,8
+accuracy,ETTm1,Time-MoE_Ultra,MAE,0.391,,5,6,9
+accuracy,MSE ↓,Toto,MAE,0.448,,5,7,2
+accuracy,MSE ↓,Moirai_Small,MAE,0.381,,5,7,3
+accuracy,MSE ↓,Moirai_Base,MAE,0.390,,5,7,4
+accuracy,MSE ↓,Moirai_Large,MAE,0.374,,5,7,5
+accuracy,MSE ↓,VisionTS,MAE,0.394,,5,7,6
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.376,,5,7,7
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.356,,5,7,8
+accuracy,ETTm2,Toto,MAE,0.303,,5,8,2
+accuracy,ETTm2,Moirai_Small,MAE,0.341,,5,8,3
+accuracy,ETTm2,Moirai_Base,MAE,0.321,,5,8,4
+accuracy,ETTm2,Moirai_Large,MAE,0.320,,5,8,5
+accuracy,ETTm2,VisionTS,MAE,0.321,,5,8,6
+accuracy,ETTm2,Time-MoE_Base,MAE,0.365,,5,8,7
+accuracy,ETTm2,Time-MoE_Large,MAE,0.361,,5,8,8
+accuracy,ETTm2,Time-MoE_Ultra,MAE,0.344,,5,8,9
+accuracy,MSE ↓,Toto,MAE,0.300,,5,9,2
+accuracy,MSE ↓,Moirai_Small,MAE,0.272,,5,9,3
+accuracy,MSE ↓,Moirai_Base,MAE,0.276,,5,9,4
+accuracy,MSE ↓,Moirai_Large,MAE,0.282,,5,9,5
+accuracy,MSE ↓,VisionTS,MAE,0.317,,5,9,6
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.316,,5,9,7
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.288,,5,9,8
+accuracy,Electricity,Toto,MAE,0.243,,5,10,2
+accuracy,Electricity,Moirai_Small,MAE,0.320,,5,10,3
+accuracy,Electricity,Moirai_Base,MAE,0.274,,5,10,4
+accuracy,Electricity,Moirai_Large,MAE,0.273,,5,10,5
+accuracy,Electricity,VisionTS,MAE,0.294,,5,10,6
+accuracy,MSE ↓,Toto,MAE,0.233,,5,11,2
+accuracy,MSE ↓,Moirai_Small,MAE,0.188,,5,11,3
+accuracy,MSE ↓,Moirai_Base,MAE,0.188,,5,11,4
+accuracy,MSE ↓,Moirai_Large,MAE,0.207,,5,11,5
+accuracy,Weather,Toto,MAE,0.245,,5,12,2
+accuracy,Weather,Moirai_Small,MAE,0.267,,5,12,3
+accuracy,Weather,Moirai_Base,MAE,0.261,,5,12,4
+accuracy,Weather,Moirai_Large,MAE,0.275,,5,12,5
+accuracy,Weather,VisionTS,MAE,0.292,,5,12,6
+accuracy,Weather,Time-MoE_Base,MAE,0.297,,5,12,7
+accuracy,Weather,Time-MoE_Large,MAE,0.300,,5,12,8
+accuracy,Weather,Time-MoE_Ultra,MAE,0.288,,5,12,9
+accuracy,MSE ↓,Toto,MAE,0.242,,5,13,2
+accuracy,MSE ↓,Moirai_Small,MAE,0.238,,5,13,3
+accuracy,MSE ↓,Moirai_Base,MAE,0.259,,5,13,4
+accuracy,MSE ↓,Moirai_Large,MAE,0.269,,5,13,5
+accuracy,MSE ↓,VisionTS,MAE,0.265,,5,13,6
+accuracy,MSE ↓,Time-MoE_Base,MAE,0.270,,5,13,7
+accuracy,MSE ↓,Time-MoE_Large,MAE,0.256,,5,13,8
+accuracy,Mean,Toto,MAE,0.324,,5,14,2
+accuracy,Mean,Moirai_Small,MAE,0.357,,5,14,3
+accuracy,Mean,Moirai_Base,MAE,0.344,,5,14,4
+accuracy,Mean,Moirai_Large,MAE,0.350,,5,14,5
+accuracy,Mean,VisionTS,MAE,0.345,,5,14,6
+accuracy,MSE ↓,Toto,MAE,0.327,,5,15,2
+accuracy,MSE ↓,Moirai_Small,MAE,0.310,,5,15,3
+accuracy,MSE ↓,Moirai_Base,MAE,0.330,,5,15,4
+accuracy,MSE ↓,Moirai_Large,MAE,0.309,,5,15,5
+accuracy,Best Count,Toto,MAE,8,,5,16,2
+accuracy,Best Count,Moirai_Small,MAE,0,,5,16,3
+accuracy,Best Count,Moirai_Base,MAE,0,,5,16,4
+accuracy,Best Count,Moirai_Large,MAE,0,,5,16,5
+accuracy,Best Count,VisionTS,MAE,3,,5,16,6
+accuracy,Best Count,Time-MoE_Base,MAE,0,,5,16,7
+accuracy,Best Count,Time-MoE_Large,MAE,0,,5,16,8
+accuracy,Best Count,Time-MoE_Ultra,MAE,1,,5,16,9
+accuracy,Count,BOOM,"MASE, CRPS",0.687,,6,1,2
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.370,,6,2,2
+accuracy,Distribution,BOOM,"MASE, CRPS",0.658,,6,3,2
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.434,,6,4,2
+accuracy,Gauge,BOOM,"MASE, CRPS",0.583,,6,5,2
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.471,,6,6,2
+accuracy,Rate,BOOM,"MASE, CRPS",0.634,,6,7,2
+accuracy,CRPS ↓,BOOM,"MASE, CRPS",0.433,,6,8,2
+accuracy,BOOM,BOOM,MASE,0.617,,7,1,2
+accuracy,CRPS $\downarrow$,BOOM,MASE,0.442,,7,2,2
+accuracy,Rank $\downarrow$,BOOM,MASE,4.905,,7,3,2
+accuracy,BOOMLET,BOOM,MASE,0.617,,7,4,2
+accuracy,CRPS $\downarrow$,BOOM,MASE,0.631,,7,5,2
+accuracy,Rank $\downarrow$,BOOM,MASE,5.711,,7,6,2
diff --git a/result/per_paper/2505.14766/components_architecture.csv b/result/per_paper/2505.14766/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..555dc8ab61a9500837fb5e6f65df5ac39793d780
--- /dev/null
+++ b/result/per_paper/2505.14766/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Decoder-Only Architecture,"A modern transformer architecture focused on time series forecasting, using only decoder layers for sequence generation.",proposed_here,,TOTO uses a modern decoder-only architecture coupled with architectural innovations designed to account for specific challenges found in multivariate observability time series data.
+Per-Variate Patch-Based Causal Scaling,A novel mechanism to address highly nonstationary sequences by scaling attention weights per variable.,proposed_here,,a novel per-variate patch-based causal scaling to address highly nonstationary sequences
+Proportional Time-Variate Factorized Attention,An attention mechanism that efficiently attends across a large number of covariates by factorizing attention over time and variables.,proposed_here,,proportional time-variate factorized attention to judiciously attend across a large number of covariates
+Student-T Mixture Prediction Head,"A probabilistic output layer using a learned Student-T Mixture model to capture complex, skewed distributions in forecasts.",proposed_here,,a Student-T mixture prediction head optimized via a robust composite loss to fit complex and highly skewed distributions
+Patch Embedding,A method to encode context time series points into embeddings for processing by attention layers.,proposed_here,,context time series points are passed through a patch embedding
diff --git a/result/per_paper/2505.14766/computational.csv b/result/per_paper/2505.14766/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..cca189956fee07b27e5ceee19fbc0a6337b101b8
--- /dev/null
+++ b/result/per_paper/2505.14766/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,No hardware type specified for BOOM model
+num_devices,,,not_reported,No device count mentioned for BOOM model
+training_cost,,,not_reported,No training cost data provided for BOOM model
+training_batch_size,,,not_reported,No batch size details for BOOM model
+training_steps_or_epochs,,,not_reported,No training duration metrics for BOOM model
+precision,,,not_reported,No precision specifications for BOOM model
+inference_latency,,,not_reported,No latency measurements for BOOM model
+inference_throughput,,,not_reported,No throughput data for BOOM model
+peak_memory,,,not_reported,No memory usage metrics for BOOM model
+flops_or_macs,,,not_reported,No computational complexity metrics for BOOM model
+num_inference_samples,,,not_reported,No sample count details for BOOM model
+params,,,not_reported,No parameter count specified for BOOM model
+context_lengths_evaluated,,,not_reported,No context length data for BOOM model
+horizon_lengths_evaluated,,,not_reported,No horizon length metrics for BOOM model
+inference_batch_size,,,not_reported,No inference batch size details for BOOM model
diff --git a/result/per_paper/2505.23719/accuracy_efficiency.csv b/result/per_paper/2505.23719/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d65cf86b1acd62bbc60bd9840b749cb85a01297b
--- /dev/null
+++ b/result/per_paper/2505.23719/accuracy_efficiency.csv
@@ -0,0 +1,333 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,bitbrains_fast_storage/5T/long,TiRex,MASE,0.655,0.028
+accuracy,bitbrains_fast_storage/5T/medium,TiRex,MASE,0.605,0.016
+accuracy,bitbrains_fast_storage/5T/short,TiRex,MASE,0.408,0.005
+accuracy,bitbrains_fast_storage/H/short,TiRex,MASE,0.699,0.013
+accuracy,bitbrains_rnd/5T/long,TiRex,MASE,0.660,0.065
+accuracy,bitbrains_rnd/5T/medium,TiRex,MASE,0.594,0.019
+accuracy,bitbrains_rnd/5T/short,TiRex,MASE,0.403,0.001
+accuracy,bitbrains_rnd/H/short,TiRex,MASE,0.631,0.013
+accuracy,bizitobs_application/10S/long,TiRex,MASE,0.053,0.004
+accuracy,bizitobs_application/10S/medium,TiRex,MASE,0.041,0.003
+accuracy,bizitobs_application/10S/short,TiRex,MASE,0.013,0.001
+accuracy,bizitobs_12c/5T/long,TiRex,MASE,0.581,0.032
+accuracy,bizitobs_12c/5T/medium,TiRex,MASE,0.366,0.015
+accuracy,bizitobs_12c/5T/short,TiRex,MASE,0.078,0.002
+accuracy,bizitobs_12c/H/long,TiRex,MASE,0.276,0.010
+accuracy,bizitobs_12c/H/medium,TiRex,MASE,0.253,0.008
+accuracy,bizitobs_12c/H/short,TiRex,MASE,0.227,0.011
+accuracy,bizitobs_service/10S/long,TiRex,MASE,0.054,0.002
+accuracy,bizitobs_service/10S/medium,TiRex,MASE,0.034,0.002
+accuracy,bizitobs_service/10S/short,TiRex,MASE,0.013,0.000
+accuracy,car_parts/M/short,TiRex,MASE,0.990,0.010
+accuracy,covid_deaths/D/short,TiRex,MASE,0.037,0.004
+accuracy,electricity/15T/long,TiRex,MASE,0.075,0.001
+accuracy,electricity/15T/medium,TiRex,MASE,0.075,0.001
+accuracy,electricity/15T/short,TiRex,MASE,0.082,0.000
+accuracy,electricity/D/short,TiRex,MASE,0.056,0.001
+accuracy,electricity/H/long,TiRex,MASE,0.092,0.003
+accuracy,electricity/H/medium,TiRex,MASE,0.078,0.002
+accuracy,electricity/H/short,TiRex,MASE,0.061,0.001
+accuracy,electricity/W/short,TiRex,MASE,0.046,0.001
+accuracy,ett1/15T/long,TiRex,MASE,0.246,0.003
+accuracy,ett1/15T/medium,TiRex,MASE,0.251,0.002
+accuracy,ett1/15T/short,TiRex,MASE,0.161,0.003
+accuracy,ett1/D/short,TiRex,MASE,0.282,0.004
+accuracy,ett1/H/long,TiRex,MASE,0.263,0.004
+accuracy,ett1/H/medium,TiRex,MASE,0.253,0.002
+accuracy,ett1/H/short,TiRex,MASE,0.179,0.002
+accuracy,ett1/W/short,TiRex,MASE,0.306,0.009
+accuracy,ett2/15T/long,TiRex,MASE,0.097,0.001
+accuracy,ett2/15T/medium,TiRex,MASE,0.093,0.001
+accuracy,ett2/15T/short,TiRex,MASE,0.066,0.001
+accuracy,ett2/D/short,TiRex,MASE,0.092,0.001
+accuracy,ett2/H/long,TiRex,MASE,0.116,0.004
+accuracy,ett2/H/medium,TiRex,MASE,0.106,0.002
+accuracy,ett2/H/short,TiRex,MASE,0.065,0.001
+accuracy,ett2/W/short,TiRex,MASE,0.088,0.002
+accuracy,hierarchical_sales/D/short,TiRex,MASE,0.572,0.002
+accuracy,hierarchical_sales/W/short,TiRex,MASE,0.349,0.003
+accuracy,hospital/M/short,TiRex,MASE,0.052,0.000
+accuracy,jena_weather/10T/long,TiRex,MASE,0.641,
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.610,
+accuracy,jena_weather/10T/short,TiRex,MASE,0.297,
+accuracy,jena_weather/D/short,TiRex,MASE,1.02,
+accuracy,jena_weather/H/long,TiRex,MASE,0.987,
+accuracy,jena_weather/H/medium,TiRex,MASE,0.828,
+accuracy,jena_weather/H/short,TiRex,MASE,0.516,
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,1.21,
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.759,
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.825,
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.657,
+accuracy,loop_seattle/5T/long,TiRex,MASE,1.02,
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.941,
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.572,
+accuracy,loop_seattle/D/short,TiRex,MASE,0.878,
+accuracy,loop_seattle/H/long,TiRex,MASE,0.917,
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.944,
+accuracy,loop_seattle/H/short,TiRex,MASE,0.850,
+accuracy,m4_daily/D/short,TiRex,MASE,3.15,
+accuracy,m4_hourly/H/short,TiRex,MASE,0.719,
+accuracy,m4_monthly/M/short,TiRex,MASE,0.929,
+accuracy,m4_quarterly/Q/short,TiRex,MASE,1.18,
+accuracy,m4_weekly/W/short,TiRex,MASE,1.90,
+accuracy,m4_yearly/A/short,TiRex,MASE,3.45,
+accuracy,m_dense/D/short,TiRex,MASE,0.688,
+accuracy,m_dense/H/long,TiRex,MASE,0.730,
+accuracy,m_dense/H/medium,TiRex,MASE,0.736,
+accuracy,m_dense/H/short,TiRex,MASE,0.788,
+accuracy,restaurant/D/short,TiRex,MASE,0.677,
+accuracy,saugeen/D/short,TiRex,MASE,3.12,
+accuracy,saugeen/M/short,TiRex,MASE,0.750,
+accuracy,saugeen/W/short,TiRex,MASE,1.18,
+accuracy,solar/10T/long,TiRex,MASE,0.828,
+accuracy,solar/10T/medium,TiRex,MASE,0.879,
+accuracy,solar/10T/short,TiRex,MASE,1.05,
+accuracy,solar/D/short,TiRex,MASE,0.971,
+accuracy,solar/H/long,TiRex,MASE,0.697,
+accuracy,solar/H/medium,TiRex,MASE,0.731,
+accuracy,solar/H/short,TiRex,MASE,0.699,
+accuracy,solar/W/short,TiRex,MASE,1.13,
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.509,
+accuracy,sz_taxi/15T/medium,TiRex,MASE,0.536,
+accuracy,sz_taxi/15T/short,TiRex,MASE,0.544,
+accuracy,sz_taxi/H/short,TiRex,MASE,0.563,
+accuracy,temperature_rain/D/short,TiRex,MASE,1.34,
+accuracy,us_births/D/short,TiRex,MASE,0.404,
+accuracy,us_births/M/short,TiRex,MASE,0.808,
+accuracy,us_births/W/short,TiRex,MASE,1.08,
+accuracy,jena_weather/10T/long,TiRex,MASE,0.053,
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.051,
+accuracy,jena_weather/10T/short,TiRex,MASE,0.030,
+accuracy,jena_weather/D/short,TiRex,MASE,0.046,
+accuracy,jena_weather/H/long,TiRex,MASE,0.057,
+accuracy,jena_weather/H/medium,TiRex,MASE,0.051,
+accuracy,jena_weather/H/short,TiRex,MASE,0.041,
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,0.381,
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.341,
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.337,
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.270,
+accuracy,loop_seattle/5T/long,TiRex,MASE,0.090,
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.083,
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.049,
+accuracy,loop_seattle/D/short,TiRex,MASE,0.042,
+accuracy,loop_seattle/H/long,TiRex,MASE,0.063,
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.065,
+accuracy,loop_seattle/H/short,TiRex,MASE,0.059,
+accuracy,m4_daily/D/short,TiRex,MASE,0.021,
+accuracy,m4_hourly/H/short,TiRex,MASE,0.021,
+accuracy,m4_monthly/M/short,TiRex,MASE,0.093,
+accuracy,m4_quarterly/Q/short,TiRex,MASE,0.074,
+accuracy,m4_weekly/W/short,TiRex,MASE,0.035,
+accuracy,m4_yearly/A/short,TiRex,MASE,0.119,
+accuracy,m_dense/D/short,TiRex,MASE,0.066,
+accuracy,m_dense/H/long,TiRex,MASE,0.122,
+accuracy,m_dense/H/medium,TiRex,MASE,0.121,
+accuracy,m_dense/H/short,TiRex,MASE,0.130,
+accuracy,restaurant/D/short,TiRex,MASE,0.254,
+accuracy,saugeen/D/short,TiRex,MASE,0.382,
+accuracy,saugeen/M/short,TiRex,MASE,0.303,
+accuracy,saugeen/W/short,TiRex,MASE,0.353,
+accuracy,solar/10T/long,TiRex,MASE,0.328,
+accuracy,solar/10T/medium,TiRex,MASE,0.354,
+accuracy,solar/10T/short,TiRex,MASE,0.542,
+accuracy,solar/D/short,TiRex,MASE,0.281,
+accuracy,solar/H/long,TiRex,MASE,0.243,
+accuracy,solar/H/medium,TiRex,MASE,0.260,
+accuracy,solar/H/short,TiRex,MASE,0.259,
+accuracy,solar/W/short,TiRex,MASE,0.154,
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.197,
+accuracy,bitbrains_fast_storage/5T/long,TiRex,MASE,0.655,0.028
+accuracy,bitbrains_fast_storage/5T/medium,TiRex,MASE,0.605,0.016
+accuracy,bitbrains_fast_storage/5T/short,TiRex,MASE,0.408,0.005
+accuracy,bitbrains_fast_storage/H/short,TiRex,MASE,0.699,0.013
+accuracy,bitbrains_rnd/5T/long,TiRex,MASE,0.660,0.065
+accuracy,bitbrains_rnd/5T/medium,TiRex,MASE,0.594,0.019
+accuracy,bitbrains_rnd/5T/short,TiRex,MASE,0.403,0.001
+accuracy,bitbrains_rnd/H/short,TiRex,MASE,0.631,0.013
+accuracy,bizitobs_application/10S/long,TiRex,MASE,0.053,0.004
+accuracy,bizitobs_application/10S/medium,TiRex,MASE,0.041,0.003
+accuracy,bizitobs_application/10S/short,TiRex,MASE,0.013,0.001
+accuracy,bizitobs_l2c/5T/long,TiRex,MASE,0.581,0.032
+accuracy,bizitobs_l2c/5T/medium,TiRex,MASE,0.366,0.015
+accuracy,bizitobs_l2c/5T/short,TiRex,MASE,0.078,0.002
+accuracy,bizitobs_l2c/H/long,TiRex,MASE,0.276,0.010
+accuracy,bizitobs_l2c/H/medium,TiRex,MASE,0.253,0.008
+accuracy,bizitobs_l2c/H/short,TiRex,MASE,0.227,0.011
+accuracy,bizitobs_service/10S/long,TiRex,MASE,0.054,0.002
+accuracy,bizitobs_service/10S/medium,TiRex,MASE,0.034,0.002
+accuracy,bizitobs_service/10S/short,TiRex,MASE,0.013,0.000
+accuracy,car_parts/M/short,TiRex,MASE,0.990,0.010
+accuracy,covid_deaths/D/short,TiRex,MASE,0.037,0.004
+accuracy,electricity/15T/long,TiRex,MASE,0.075,0.001
+accuracy,electricity/15T/medium,TiRex,MASE,0.075,0.001
+accuracy,electricity/15T/short,TiRex,MASE,0.082,0.000
+accuracy,electricity/D/short,TiRex,MASE,0.056,0.001
+accuracy,electricity/H/long,TiRex,MASE,0.092,0.003
+accuracy,electricity/H/medium,TiRex,MASE,0.078,0.002
+accuracy,electricity/H/short,TiRex,MASE,0.061,0.001
+accuracy,electricity/W/short,TiRex,MASE,0.046,0.001
+accuracy,ett1/15T/long,TiRex,MASE,0.246,0.003
+accuracy,ett1/15T/medium,TiRex,MASE,0.251,0.002
+accuracy,ett1/15T/short,TiRex,MASE,0.161,0.003
+accuracy,ett1/D/short,TiRex,MASE,0.282,0.004
+accuracy,ett1/H/long,TiRex,MASE,0.263,0.004
+accuracy,ett1/H/medium,TiRex,MASE,0.253,0.002
+accuracy,ett1/H/short,TiRex,MASE,0.179,0.002
+accuracy,ett1/W/short,TiRex,MASE,0.306,0.009
+accuracy,ett2/15T/long,TiRex,MASE,0.097,0.001
+accuracy,ett2/15T/medium,TiRex,MASE,0.093,0.001
+accuracy,ett2/15T/short,TiRex,MASE,0.066,0.001
+accuracy,ett2/D/short,TiRex,MASE,0.092,0.001
+accuracy,ett2/H/long,TiRex,MASE,0.116,0.004
+accuracy,ett2/H/medium,TiRex,MASE,0.106,0.002
+accuracy,ett2/H/short,TiRex,MASE,0.065,0.001
+accuracy,ett2/W/short,TiRex,MASE,0.088,0.002
+accuracy,hierarchical_sales/D/short,TiRex,MASE,0.572,0.002
+accuracy,hierarchical_sales/W/short,TiRex,MASE,0.349,0.003
+accuracy,hospital/M/short,TiRex,MASE,0.052,0.000
+accuracy,bitbrains_fast_storage/5T/long,TiRex,MASE,0.916,0.006
+accuracy,bitbrains_fast_storage/5T/medium,TiRex,MASE,1.00,0.011
+accuracy,bitbrains_fast_storage/5T/short,TiRex,MASE,0.692,0.007
+accuracy,bitbrains_fast_storage/H/short,TiRex,MASE,1.06,0.013
+accuracy,bitbrains_rnd/5T/long,TiRex,MASE,3.35,0.013
+accuracy,bitbrains_rnd/5T/medium,TiRex,MASE,4.40,0.007
+accuracy,bitbrains_rnd/5T/short,TiRex,MASE,1.66,0.004
+accuracy,bitbrains_rnd/H/short,TiRex,MASE,5.84,0.011
+accuracy,bizitobs_application/10S/long,TiRex,MASE,3.67,0.069
+accuracy,bizitobs_application/10S/medium,TiRex,MASE,2.85,0.057
+accuracy,bizitobs_application/10S/short,TiRex,MASE,1.40,0.112
+accuracy,bizitobs_12c/5T/long,TiRex,MASE,1.19,0.040
+accuracy,bizitobs_12c/5T/medium,TiRex,MASE,0.849,0.028
+accuracy,bizitobs_12c/5T/short,TiRex,MASE,0.290,0.005
+accuracy,bizitobs_12c/H/long,TiRex,MASE,0.590,0.022
+accuracy,bizitobs_12c/H/medium,TiRex,MASE,0.525,0.020
+accuracy,bizitobs_12c/H/short,TiRex,MASE,0.528,0.021
+accuracy,bizitobs_service/10S/long,TiRex,MASE,1.57,0.057
+accuracy,bizitobs_service/10S/medium,TiRex,MASE,1.32,0.058
+accuracy,bizitobs_service/10S/short,TiRex,MASE,0.884,0.052
+accuracy,car_parts/M/short,TiRex,MASE,0.838,0.004
+accuracy,covid_deaths/D/short,TiRex,MASE,39.5,0.803
+accuracy,electricity/15T/long,TiRex,MASE,0.891,0.008
+accuracy,electricity/15T/medium,TiRex,MASE,0.841,0.006
+accuracy,electricity/15T/short,TiRex,MASE,0.945,0.008
+accuracy,electricity/D/short,TiRex,MASE,1.43,0.010
+accuracy,electricity/H/long,TiRex,MASE,1.21,0.018
+accuracy,electricity/H/medium,TiRex,MASE,1.08,0.010
+accuracy,electricity/H/short,TiRex,MASE,0.869,0.009
+accuracy,electricity/W/short,TiRex,MASE,1.46,0.010
+accuracy,ett1/15T/long,TiRex,MASE,1.05,0.009
+accuracy,ett1/15T/medium,TiRex,MASE,1.04,0.009
+accuracy,ett1/15T/short,TiRex,MASE,0.706,0.007
+accuracy,ett1/D/short,TiRex,MASE,1.71,0.016
+accuracy,ett1/H/long,TiRex,MASE,1.34,0.030
+accuracy,ett1/H/medium,TiRex,MASE,1.25,0.017
+accuracy,ett1/H/short,TiRex,MASE,0.827,0.007
+accuracy,ett1/W/short,TiRex,MASE,1.72,0.044
+accuracy,ett2/15T/long,TiRex,MASE,0.932,0.012
+accuracy,ett2/15T/medium,TiRex,MASE,0.910,0.010
+accuracy,ett2/15T/short,TiRex,MASE,0.749,0.010
+accuracy,ett2/D/short,TiRex,MASE,1.28,0.019
+accuracy,ett2/H/long,TiRex,MASE,1.16,0.037
+accuracy,ett2/H/medium,TiRex,MASE,1.05,0.021
+accuracy,ett2/H/short,TiRex,MASE,0.742,0.006
+accuracy,ett2/W/short,TiRex,MASE,0.797,0.040
+accuracy,hierarchical_sales/D/short,TiRex,MASE,0.744,0.002
+accuracy,hierarchical_sales/W/short,TiRex,MASE,0.721,0.001
+accuracy,hospital/M/short,TiRex,MASE,0.767,0.003
+accuracy,jena_weather/10T/long,TiRex,MASE,0.641,
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.610,
+accuracy,jena_weather/10T/short,TiRex,MASE,0.297,
+accuracy,jena_weather/D/short,TiRex,MASE,1.02,
+accuracy,jena_weather/H/long,TiRex,MASE,0.987,
+accuracy,jena_weather/H/medium,TiRex,MASE,0.828,
+accuracy,jena_weather/H/short,TiRex,MASE,0.516,
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,1.21,
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.759,
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.825,
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.657,
+accuracy,loop_seattle/5T/long,TiRex,MASE,1.02,
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.941,
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.572,
+accuracy,loop_seattle/D/short,TiRex,MASE,0.878,
+accuracy,loop_seattle/H/long,TiRex,MASE,0.917,
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.944,
+accuracy,loop_seattle/H/short,TiRex,MASE,0.850,
+accuracy,m4_daily/D/short,TiRex,MASE,3.15,
+accuracy,m4_hourly/H/short,TiRex,MASE,0.719,
+accuracy,m4_monthly/M/short,TiRex,MASE,0.929,
+accuracy,m4_quarterly/Q/short,TiRex,MASE,1.18,
+accuracy,m4_weekly/W/short,TiRex,MASE,1.90,
+accuracy,m4_yearly/A/short,TiRex,MASE,3.45,
+accuracy,m_dense/D/short,TiRex,MASE,0.688,
+accuracy,m_dense/H/long,TiRex,MASE,0.730,
+accuracy,m_dense/H/medium,TiRex,MASE,0.736,
+accuracy,m_dense/H/short,TiRex,MASE,0.788,
+accuracy,restaurant/D/short,TiRex,MASE,0.677,
+accuracy,saugeen/D/short,TiRex,MASE,3.12,
+accuracy,saugeen/M/short,TiRex,MASE,0.750,
+accuracy,saugeen/W/short,TiRex,MASE,1.18,
+accuracy,solar/10T/long,TiRex,MASE,0.828,
+accuracy,solar/10T/medium,TiRex,MASE,0.879,
+accuracy,solar/10T/short,TiRex,MASE,1.05,
+accuracy,solar/D/short,TiRex,MASE,0.971,
+accuracy,solar/H/long,TiRex,MASE,0.697,
+accuracy,solar/H/medium,TiRex,MASE,0.731,
+accuracy,solar/H/short,TiRex,MASE,0.699,
+accuracy,solar/W/short,TiRex,MASE,1.13,
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.509,
+accuracy,sz_taxi/15T/medium,TiRex,MASE,0.536,
+accuracy,sz_taxi/15T/short,TiRex,MASE,0.544,
+accuracy,sz_taxi/H/short,TiRex,MASE,0.563,
+accuracy,temperature_rain/D/short,TiRex,MASE,1.34,
+accuracy,us_births/D/short,TiRex,MASE,0.404,
+accuracy,us_births/M/short,TiRex,MASE,0.808,
+accuracy,us_births/W/short,TiRex,MASE,1.08,
+accuracy,jena_weather/10T/long,TiRex,MASE,0.053,
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.051,
+accuracy,jena_weather/10T/short,TiRex,MASE,0.030,
+accuracy,jena_weather/D/short,TiRex,MASE,0.046,
+accuracy,jena_weather/H/long,TiRex,MASE,0.057,
+accuracy,jena_weather/H/medium,TiRex,MASE,0.051,
+accuracy,jena_weather/H/short,TiRex,MASE,0.041,
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,0.381,
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.341,
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.337,
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.270,
+accuracy,loop_seattle/5T/long,TiRex,MASE,0.090,
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.083,
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.049,
+accuracy,loop_seattle/D/short,TiRex,MASE,0.042,
+accuracy,loop_seattle/H/long,TiRex,MASE,0.063,
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.065,
+accuracy,loop_seattle/H/short,TiRex,MASE,0.059,
+accuracy,m4_daily/D/short,TiRex,MASE,0.021,
+accuracy,m4_hourly/H/short,TiRex,MASE,0.021,
+accuracy,m4_monthly/M/short,TiRex,MASE,0.093,
+accuracy,m4_quarterly/Q/short,TiRex,MASE,0.074,
+accuracy,m4_weekly/W/short,TiRex,MASE,0.035,
+accuracy,m4_yearly/A/short,TiRex,MASE,0.119,
+accuracy,m_dense/D/short,TiRex,MASE,0.066,
+accuracy,m_dense/H/long,TiRex,MASE,0.122,
+accuracy,m_dense/H/medium,TiRex,MASE,0.121,
+accuracy,m_dense/H/short,TiRex,MASE,0.130,
+accuracy,restaurant/D/short,TiRex,MASE,0.254,
+accuracy,saugeen/D/short,TiRex,MASE,0.382,
+accuracy,saugeen/M/short,TiRex,MASE,0.303,
+accuracy,saugeen/W/short,TiRex,MASE,0.353,
+accuracy,solar/10T/long,TiRex,MASE,0.328,
+accuracy,solar/10T/medium,TiRex,MASE,0.354,
+accuracy,solar/10T/short,TiRex,MASE,0.542,
+accuracy,solar/D/short,TiRex,MASE,0.281,
+accuracy,solar/H/long,TiRex,MASE,0.243,
+accuracy,solar/H/medium,TiRex,MASE,0.260,
+accuracy,solar/H/short,TiRex,MASE,0.259,
+accuracy,solar/W/short,TiRex,MASE,0.154,
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.197,
+accuracy,sz_taxi/15T/medium,TiRex,MASE,0.202,
+accuracy,sz_taxi/15T/short,TiRex,MASE,0.200,
+accuracy,sz_taxi/H/short,TiRex,MASE,0.136,
+accuracy,temperature_rain/D/short,TiRex,MASE,0.550,
+accuracy,us_births/D/short,TiRex,MASE,0.021,
+accuracy,us_births/M/short,TiRex,MASE,0.017,
+accuracy,us_births/W/short,TiRex,MASE,0.013,
diff --git a/result/per_paper/2505.23719/accuracy_efficiency_traced.csv b/result/per_paper/2505.23719/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..dd2cfb7b956c683d4574c9f6729e6abc81eadec6
--- /dev/null
+++ b/result/per_paper/2505.23719/accuracy_efficiency_traced.csv
@@ -0,0 +1,333 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,bitbrains_fast_storage/5T/long,TiRex,MASE,0.655,0.028,0,1,1
+accuracy,bitbrains_fast_storage/5T/medium,TiRex,MASE,0.605,0.016,0,2,1
+accuracy,bitbrains_fast_storage/5T/short,TiRex,MASE,0.408,0.005,0,3,1
+accuracy,bitbrains_fast_storage/H/short,TiRex,MASE,0.699,0.013,0,4,1
+accuracy,bitbrains_rnd/5T/long,TiRex,MASE,0.660,0.065,0,5,1
+accuracy,bitbrains_rnd/5T/medium,TiRex,MASE,0.594,0.019,0,6,1
+accuracy,bitbrains_rnd/5T/short,TiRex,MASE,0.403,0.001,0,7,1
+accuracy,bitbrains_rnd/H/short,TiRex,MASE,0.631,0.013,0,8,1
+accuracy,bizitobs_application/10S/long,TiRex,MASE,0.053,0.004,0,9,1
+accuracy,bizitobs_application/10S/medium,TiRex,MASE,0.041,0.003,0,10,1
+accuracy,bizitobs_application/10S/short,TiRex,MASE,0.013,0.001,0,11,1
+accuracy,bizitobs_12c/5T/long,TiRex,MASE,0.581,0.032,0,12,1
+accuracy,bizitobs_12c/5T/medium,TiRex,MASE,0.366,0.015,0,13,1
+accuracy,bizitobs_12c/5T/short,TiRex,MASE,0.078,0.002,0,14,1
+accuracy,bizitobs_12c/H/long,TiRex,MASE,0.276,0.010,0,15,1
+accuracy,bizitobs_12c/H/medium,TiRex,MASE,0.253,0.008,0,16,1
+accuracy,bizitobs_12c/H/short,TiRex,MASE,0.227,0.011,0,17,1
+accuracy,bizitobs_service/10S/long,TiRex,MASE,0.054,0.002,0,18,1
+accuracy,bizitobs_service/10S/medium,TiRex,MASE,0.034,0.002,0,19,1
+accuracy,bizitobs_service/10S/short,TiRex,MASE,0.013,0.000,0,20,1
+accuracy,car_parts/M/short,TiRex,MASE,0.990,0.010,0,21,1
+accuracy,covid_deaths/D/short,TiRex,MASE,0.037,0.004,0,22,1
+accuracy,electricity/15T/long,TiRex,MASE,0.075,0.001,0,23,1
+accuracy,electricity/15T/medium,TiRex,MASE,0.075,0.001,0,24,1
+accuracy,electricity/15T/short,TiRex,MASE,0.082,0.000,0,25,1
+accuracy,electricity/D/short,TiRex,MASE,0.056,0.001,0,26,1
+accuracy,electricity/H/long,TiRex,MASE,0.092,0.003,0,27,1
+accuracy,electricity/H/medium,TiRex,MASE,0.078,0.002,0,28,1
+accuracy,electricity/H/short,TiRex,MASE,0.061,0.001,0,29,1
+accuracy,electricity/W/short,TiRex,MASE,0.046,0.001,0,30,1
+accuracy,ett1/15T/long,TiRex,MASE,0.246,0.003,0,31,1
+accuracy,ett1/15T/medium,TiRex,MASE,0.251,0.002,0,32,1
+accuracy,ett1/15T/short,TiRex,MASE,0.161,0.003,0,33,1
+accuracy,ett1/D/short,TiRex,MASE,0.282,0.004,0,34,1
+accuracy,ett1/H/long,TiRex,MASE,0.263,0.004,0,35,1
+accuracy,ett1/H/medium,TiRex,MASE,0.253,0.002,0,36,1
+accuracy,ett1/H/short,TiRex,MASE,0.179,0.002,0,37,1
+accuracy,ett1/W/short,TiRex,MASE,0.306,0.009,0,38,1
+accuracy,ett2/15T/long,TiRex,MASE,0.097,0.001,0,39,1
+accuracy,ett2/15T/medium,TiRex,MASE,0.093,0.001,0,40,1
+accuracy,ett2/15T/short,TiRex,MASE,0.066,0.001,0,41,1
+accuracy,ett2/D/short,TiRex,MASE,0.092,0.001,0,42,1
+accuracy,ett2/H/long,TiRex,MASE,0.116,0.004,0,43,1
+accuracy,ett2/H/medium,TiRex,MASE,0.106,0.002,0,44,1
+accuracy,ett2/H/short,TiRex,MASE,0.065,0.001,0,45,1
+accuracy,ett2/W/short,TiRex,MASE,0.088,0.002,0,46,1
+accuracy,hierarchical_sales/D/short,TiRex,MASE,0.572,0.002,0,47,1
+accuracy,hierarchical_sales/W/short,TiRex,MASE,0.349,0.003,0,48,1
+accuracy,hospital/M/short,TiRex,MASE,0.052,0.000,0,49,1
+accuracy,jena_weather/10T/long,TiRex,MASE,0.641,,1,1,1
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.610,,1,2,1
+accuracy,jena_weather/10T/short,TiRex,MASE,0.297,,1,3,1
+accuracy,jena_weather/D/short,TiRex,MASE,1.02,,1,4,1
+accuracy,jena_weather/H/long,TiRex,MASE,0.987,,1,5,1
+accuracy,jena_weather/H/medium,TiRex,MASE,0.828,,1,6,1
+accuracy,jena_weather/H/short,TiRex,MASE,0.516,,1,7,1
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,1.21,,1,8,1
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.759,,1,9,1
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.825,,1,10,1
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.657,,1,11,1
+accuracy,loop_seattle/5T/long,TiRex,MASE,1.02,,1,12,1
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.941,,1,13,1
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.572,,1,14,1
+accuracy,loop_seattle/D/short,TiRex,MASE,0.878,,1,15,1
+accuracy,loop_seattle/H/long,TiRex,MASE,0.917,,1,16,1
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.944,,1,17,1
+accuracy,loop_seattle/H/short,TiRex,MASE,0.850,,1,18,1
+accuracy,m4_daily/D/short,TiRex,MASE,3.15,,1,19,1
+accuracy,m4_hourly/H/short,TiRex,MASE,0.719,,1,20,1
+accuracy,m4_monthly/M/short,TiRex,MASE,0.929,,1,21,1
+accuracy,m4_quarterly/Q/short,TiRex,MASE,1.18,,1,22,1
+accuracy,m4_weekly/W/short,TiRex,MASE,1.90,,1,23,1
+accuracy,m4_yearly/A/short,TiRex,MASE,3.45,,1,24,1
+accuracy,m_dense/D/short,TiRex,MASE,0.688,,1,25,1
+accuracy,m_dense/H/long,TiRex,MASE,0.730,,1,26,1
+accuracy,m_dense/H/medium,TiRex,MASE,0.736,,1,27,1
+accuracy,m_dense/H/short,TiRex,MASE,0.788,,1,28,1
+accuracy,restaurant/D/short,TiRex,MASE,0.677,,1,29,1
+accuracy,saugeen/D/short,TiRex,MASE,3.12,,1,30,1
+accuracy,saugeen/M/short,TiRex,MASE,0.750,,1,31,1
+accuracy,saugeen/W/short,TiRex,MASE,1.18,,1,32,1
+accuracy,solar/10T/long,TiRex,MASE,0.828,,1,33,1
+accuracy,solar/10T/medium,TiRex,MASE,0.879,,1,34,1
+accuracy,solar/10T/short,TiRex,MASE,1.05,,1,35,1
+accuracy,solar/D/short,TiRex,MASE,0.971,,1,36,1
+accuracy,solar/H/long,TiRex,MASE,0.697,,1,37,1
+accuracy,solar/H/medium,TiRex,MASE,0.731,,1,38,1
+accuracy,solar/H/short,TiRex,MASE,0.699,,1,39,1
+accuracy,solar/W/short,TiRex,MASE,1.13,,1,40,1
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.509,,1,41,1
+accuracy,sz_taxi/15T/medium,TiRex,MASE,0.536,,1,42,1
+accuracy,sz_taxi/15T/short,TiRex,MASE,0.544,,1,43,1
+accuracy,sz_taxi/H/short,TiRex,MASE,0.563,,1,44,1
+accuracy,temperature_rain/D/short,TiRex,MASE,1.34,,1,45,1
+accuracy,us_births/D/short,TiRex,MASE,0.404,,1,46,1
+accuracy,us_births/M/short,TiRex,MASE,0.808,,1,47,1
+accuracy,us_births/W/short,TiRex,MASE,1.08,,1,48,1
+accuracy,jena_weather/10T/long,TiRex,MASE,0.053,,2,1,1
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.051,,2,2,1
+accuracy,jena_weather/10T/short,TiRex,MASE,0.030,,2,3,1
+accuracy,jena_weather/D/short,TiRex,MASE,0.046,,2,4,1
+accuracy,jena_weather/H/long,TiRex,MASE,0.057,,2,5,1
+accuracy,jena_weather/H/medium,TiRex,MASE,0.051,,2,6,1
+accuracy,jena_weather/H/short,TiRex,MASE,0.041,,2,7,1
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,0.381,,2,8,1
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.341,,2,9,1
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.337,,2,10,1
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.270,,2,11,1
+accuracy,loop_seattle/5T/long,TiRex,MASE,0.090,,2,12,1
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.083,,2,13,1
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.049,,2,14,1
+accuracy,loop_seattle/D/short,TiRex,MASE,0.042,,2,15,1
+accuracy,loop_seattle/H/long,TiRex,MASE,0.063,,2,16,1
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.065,,2,17,1
+accuracy,loop_seattle/H/short,TiRex,MASE,0.059,,2,18,1
+accuracy,m4_daily/D/short,TiRex,MASE,0.021,,2,19,1
+accuracy,m4_hourly/H/short,TiRex,MASE,0.021,,2,20,1
+accuracy,m4_monthly/M/short,TiRex,MASE,0.093,,2,21,1
+accuracy,m4_quarterly/Q/short,TiRex,MASE,0.074,,2,22,1
+accuracy,m4_weekly/W/short,TiRex,MASE,0.035,,2,23,1
+accuracy,m4_yearly/A/short,TiRex,MASE,0.119,,2,24,1
+accuracy,m_dense/D/short,TiRex,MASE,0.066,,2,25,1
+accuracy,m_dense/H/long,TiRex,MASE,0.122,,2,26,1
+accuracy,m_dense/H/medium,TiRex,MASE,0.121,,2,27,1
+accuracy,m_dense/H/short,TiRex,MASE,0.130,,2,28,1
+accuracy,restaurant/D/short,TiRex,MASE,0.254,,2,29,1
+accuracy,saugeen/D/short,TiRex,MASE,0.382,,2,30,1
+accuracy,saugeen/M/short,TiRex,MASE,0.303,,2,31,1
+accuracy,saugeen/W/short,TiRex,MASE,0.353,,2,32,1
+accuracy,solar/10T/long,TiRex,MASE,0.328,,2,33,1
+accuracy,solar/10T/medium,TiRex,MASE,0.354,,2,34,1
+accuracy,solar/10T/short,TiRex,MASE,0.542,,2,35,1
+accuracy,solar/D/short,TiRex,MASE,0.281,,2,36,1
+accuracy,solar/H/long,TiRex,MASE,0.243,,2,37,1
+accuracy,solar/H/medium,TiRex,MASE,0.260,,2,38,1
+accuracy,solar/H/short,TiRex,MASE,0.259,,2,39,1
+accuracy,solar/W/short,TiRex,MASE,0.154,,2,40,1
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.197,,2,41,1
+accuracy,bitbrains_fast_storage/5T/long,TiRex,MASE,0.655,0.028,3,1,1
+accuracy,bitbrains_fast_storage/5T/medium,TiRex,MASE,0.605,0.016,3,2,1
+accuracy,bitbrains_fast_storage/5T/short,TiRex,MASE,0.408,0.005,3,3,1
+accuracy,bitbrains_fast_storage/H/short,TiRex,MASE,0.699,0.013,3,4,1
+accuracy,bitbrains_rnd/5T/long,TiRex,MASE,0.660,0.065,3,5,1
+accuracy,bitbrains_rnd/5T/medium,TiRex,MASE,0.594,0.019,3,6,1
+accuracy,bitbrains_rnd/5T/short,TiRex,MASE,0.403,0.001,3,7,1
+accuracy,bitbrains_rnd/H/short,TiRex,MASE,0.631,0.013,3,8,1
+accuracy,bizitobs_application/10S/long,TiRex,MASE,0.053,0.004,3,9,1
+accuracy,bizitobs_application/10S/medium,TiRex,MASE,0.041,0.003,3,10,1
+accuracy,bizitobs_application/10S/short,TiRex,MASE,0.013,0.001,3,11,1
+accuracy,bizitobs_l2c/5T/long,TiRex,MASE,0.581,0.032,3,12,1
+accuracy,bizitobs_l2c/5T/medium,TiRex,MASE,0.366,0.015,3,13,1
+accuracy,bizitobs_l2c/5T/short,TiRex,MASE,0.078,0.002,3,14,1
+accuracy,bizitobs_l2c/H/long,TiRex,MASE,0.276,0.010,3,15,1
+accuracy,bizitobs_l2c/H/medium,TiRex,MASE,0.253,0.008,3,16,1
+accuracy,bizitobs_l2c/H/short,TiRex,MASE,0.227,0.011,3,17,1
+accuracy,bizitobs_service/10S/long,TiRex,MASE,0.054,0.002,3,18,1
+accuracy,bizitobs_service/10S/medium,TiRex,MASE,0.034,0.002,3,19,1
+accuracy,bizitobs_service/10S/short,TiRex,MASE,0.013,0.000,3,20,1
+accuracy,car_parts/M/short,TiRex,MASE,0.990,0.010,3,21,1
+accuracy,covid_deaths/D/short,TiRex,MASE,0.037,0.004,3,22,1
+accuracy,electricity/15T/long,TiRex,MASE,0.075,0.001,3,23,1
+accuracy,electricity/15T/medium,TiRex,MASE,0.075,0.001,3,24,1
+accuracy,electricity/15T/short,TiRex,MASE,0.082,0.000,3,25,1
+accuracy,electricity/D/short,TiRex,MASE,0.056,0.001,3,26,1
+accuracy,electricity/H/long,TiRex,MASE,0.092,0.003,3,27,1
+accuracy,electricity/H/medium,TiRex,MASE,0.078,0.002,3,28,1
+accuracy,electricity/H/short,TiRex,MASE,0.061,0.001,3,29,1
+accuracy,electricity/W/short,TiRex,MASE,0.046,0.001,3,30,1
+accuracy,ett1/15T/long,TiRex,MASE,0.246,0.003,3,31,1
+accuracy,ett1/15T/medium,TiRex,MASE,0.251,0.002,3,32,1
+accuracy,ett1/15T/short,TiRex,MASE,0.161,0.003,3,33,1
+accuracy,ett1/D/short,TiRex,MASE,0.282,0.004,3,34,1
+accuracy,ett1/H/long,TiRex,MASE,0.263,0.004,3,35,1
+accuracy,ett1/H/medium,TiRex,MASE,0.253,0.002,3,36,1
+accuracy,ett1/H/short,TiRex,MASE,0.179,0.002,3,37,1
+accuracy,ett1/W/short,TiRex,MASE,0.306,0.009,3,38,1
+accuracy,ett2/15T/long,TiRex,MASE,0.097,0.001,3,39,1
+accuracy,ett2/15T/medium,TiRex,MASE,0.093,0.001,3,40,1
+accuracy,ett2/15T/short,TiRex,MASE,0.066,0.001,3,41,1
+accuracy,ett2/D/short,TiRex,MASE,0.092,0.001,3,42,1
+accuracy,ett2/H/long,TiRex,MASE,0.116,0.004,3,43,1
+accuracy,ett2/H/medium,TiRex,MASE,0.106,0.002,3,44,1
+accuracy,ett2/H/short,TiRex,MASE,0.065,0.001,3,45,1
+accuracy,ett2/W/short,TiRex,MASE,0.088,0.002,3,46,1
+accuracy,hierarchical_sales/D/short,TiRex,MASE,0.572,0.002,3,47,1
+accuracy,hierarchical_sales/W/short,TiRex,MASE,0.349,0.003,3,48,1
+accuracy,hospital/M/short,TiRex,MASE,0.052,0.000,3,49,1
+accuracy,bitbrains_fast_storage/5T/long,TiRex,MASE,0.916,0.006,4,1,1
+accuracy,bitbrains_fast_storage/5T/medium,TiRex,MASE,1.00,0.011,4,2,1
+accuracy,bitbrains_fast_storage/5T/short,TiRex,MASE,0.692,0.007,4,3,1
+accuracy,bitbrains_fast_storage/H/short,TiRex,MASE,1.06,0.013,4,4,1
+accuracy,bitbrains_rnd/5T/long,TiRex,MASE,3.35,0.013,4,5,1
+accuracy,bitbrains_rnd/5T/medium,TiRex,MASE,4.40,0.007,4,6,1
+accuracy,bitbrains_rnd/5T/short,TiRex,MASE,1.66,0.004,4,7,1
+accuracy,bitbrains_rnd/H/short,TiRex,MASE,5.84,0.011,4,8,1
+accuracy,bizitobs_application/10S/long,TiRex,MASE,3.67,0.069,4,9,1
+accuracy,bizitobs_application/10S/medium,TiRex,MASE,2.85,0.057,4,10,1
+accuracy,bizitobs_application/10S/short,TiRex,MASE,1.40,0.112,4,11,1
+accuracy,bizitobs_12c/5T/long,TiRex,MASE,1.19,0.040,4,12,1
+accuracy,bizitobs_12c/5T/medium,TiRex,MASE,0.849,0.028,4,13,1
+accuracy,bizitobs_12c/5T/short,TiRex,MASE,0.290,0.005,4,14,1
+accuracy,bizitobs_12c/H/long,TiRex,MASE,0.590,0.022,4,15,1
+accuracy,bizitobs_12c/H/medium,TiRex,MASE,0.525,0.020,4,16,1
+accuracy,bizitobs_12c/H/short,TiRex,MASE,0.528,0.021,4,17,1
+accuracy,bizitobs_service/10S/long,TiRex,MASE,1.57,0.057,4,18,1
+accuracy,bizitobs_service/10S/medium,TiRex,MASE,1.32,0.058,4,19,1
+accuracy,bizitobs_service/10S/short,TiRex,MASE,0.884,0.052,4,20,1
+accuracy,car_parts/M/short,TiRex,MASE,0.838,0.004,4,21,1
+accuracy,covid_deaths/D/short,TiRex,MASE,39.5,0.803,4,22,1
+accuracy,electricity/15T/long,TiRex,MASE,0.891,0.008,4,23,1
+accuracy,electricity/15T/medium,TiRex,MASE,0.841,0.006,4,24,1
+accuracy,electricity/15T/short,TiRex,MASE,0.945,0.008,4,25,1
+accuracy,electricity/D/short,TiRex,MASE,1.43,0.010,4,26,1
+accuracy,electricity/H/long,TiRex,MASE,1.21,0.018,4,27,1
+accuracy,electricity/H/medium,TiRex,MASE,1.08,0.010,4,28,1
+accuracy,electricity/H/short,TiRex,MASE,0.869,0.009,4,29,1
+accuracy,electricity/W/short,TiRex,MASE,1.46,0.010,4,30,1
+accuracy,ett1/15T/long,TiRex,MASE,1.05,0.009,4,31,1
+accuracy,ett1/15T/medium,TiRex,MASE,1.04,0.009,4,32,1
+accuracy,ett1/15T/short,TiRex,MASE,0.706,0.007,4,33,1
+accuracy,ett1/D/short,TiRex,MASE,1.71,0.016,4,34,1
+accuracy,ett1/H/long,TiRex,MASE,1.34,0.030,4,35,1
+accuracy,ett1/H/medium,TiRex,MASE,1.25,0.017,4,36,1
+accuracy,ett1/H/short,TiRex,MASE,0.827,0.007,4,37,1
+accuracy,ett1/W/short,TiRex,MASE,1.72,0.044,4,38,1
+accuracy,ett2/15T/long,TiRex,MASE,0.932,0.012,4,39,1
+accuracy,ett2/15T/medium,TiRex,MASE,0.910,0.010,4,40,1
+accuracy,ett2/15T/short,TiRex,MASE,0.749,0.010,4,41,1
+accuracy,ett2/D/short,TiRex,MASE,1.28,0.019,4,42,1
+accuracy,ett2/H/long,TiRex,MASE,1.16,0.037,4,43,1
+accuracy,ett2/H/medium,TiRex,MASE,1.05,0.021,4,44,1
+accuracy,ett2/H/short,TiRex,MASE,0.742,0.006,4,45,1
+accuracy,ett2/W/short,TiRex,MASE,0.797,0.040,4,46,1
+accuracy,hierarchical_sales/D/short,TiRex,MASE,0.744,0.002,4,47,1
+accuracy,hierarchical_sales/W/short,TiRex,MASE,0.721,0.001,4,48,1
+accuracy,hospital/M/short,TiRex,MASE,0.767,0.003,4,49,1
+accuracy,jena_weather/10T/long,TiRex,MASE,0.641,,6,1,1
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.610,,6,2,1
+accuracy,jena_weather/10T/short,TiRex,MASE,0.297,,6,3,1
+accuracy,jena_weather/D/short,TiRex,MASE,1.02,,6,4,1
+accuracy,jena_weather/H/long,TiRex,MASE,0.987,,6,5,1
+accuracy,jena_weather/H/medium,TiRex,MASE,0.828,,6,6,1
+accuracy,jena_weather/H/short,TiRex,MASE,0.516,,6,7,1
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,1.21,,6,8,1
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.759,,6,9,1
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.825,,6,10,1
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.657,,6,11,1
+accuracy,loop_seattle/5T/long,TiRex,MASE,1.02,,6,12,1
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.941,,6,13,1
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.572,,6,14,1
+accuracy,loop_seattle/D/short,TiRex,MASE,0.878,,6,15,1
+accuracy,loop_seattle/H/long,TiRex,MASE,0.917,,6,16,1
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.944,,6,17,1
+accuracy,loop_seattle/H/short,TiRex,MASE,0.850,,6,18,1
+accuracy,m4_daily/D/short,TiRex,MASE,3.15,,6,19,1
+accuracy,m4_hourly/H/short,TiRex,MASE,0.719,,6,20,1
+accuracy,m4_monthly/M/short,TiRex,MASE,0.929,,6,21,1
+accuracy,m4_quarterly/Q/short,TiRex,MASE,1.18,,6,22,1
+accuracy,m4_weekly/W/short,TiRex,MASE,1.90,,6,23,1
+accuracy,m4_yearly/A/short,TiRex,MASE,3.45,,6,24,1
+accuracy,m_dense/D/short,TiRex,MASE,0.688,,6,25,1
+accuracy,m_dense/H/long,TiRex,MASE,0.730,,6,26,1
+accuracy,m_dense/H/medium,TiRex,MASE,0.736,,6,27,1
+accuracy,m_dense/H/short,TiRex,MASE,0.788,,6,28,1
+accuracy,restaurant/D/short,TiRex,MASE,0.677,,6,29,1
+accuracy,saugeen/D/short,TiRex,MASE,3.12,,6,30,1
+accuracy,saugeen/M/short,TiRex,MASE,0.750,,6,31,1
+accuracy,saugeen/W/short,TiRex,MASE,1.18,,6,32,1
+accuracy,solar/10T/long,TiRex,MASE,0.828,,6,33,1
+accuracy,solar/10T/medium,TiRex,MASE,0.879,,6,34,1
+accuracy,solar/10T/short,TiRex,MASE,1.05,,6,35,1
+accuracy,solar/D/short,TiRex,MASE,0.971,,6,36,1
+accuracy,solar/H/long,TiRex,MASE,0.697,,6,37,1
+accuracy,solar/H/medium,TiRex,MASE,0.731,,6,38,1
+accuracy,solar/H/short,TiRex,MASE,0.699,,6,39,1
+accuracy,solar/W/short,TiRex,MASE,1.13,,6,40,1
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.509,,6,41,1
+accuracy,sz_taxi/15T/medium,TiRex,MASE,0.536,,6,42,1
+accuracy,sz_taxi/15T/short,TiRex,MASE,0.544,,6,43,1
+accuracy,sz_taxi/H/short,TiRex,MASE,0.563,,6,44,1
+accuracy,temperature_rain/D/short,TiRex,MASE,1.34,,6,45,1
+accuracy,us_births/D/short,TiRex,MASE,0.404,,6,46,1
+accuracy,us_births/M/short,TiRex,MASE,0.808,,6,47,1
+accuracy,us_births/W/short,TiRex,MASE,1.08,,6,48,1
+accuracy,jena_weather/10T/long,TiRex,MASE,0.053,,7,1,1
+accuracy,jena_weather/10T/medium,TiRex,MASE,0.051,,7,2,1
+accuracy,jena_weather/10T/short,TiRex,MASE,0.030,,7,3,1
+accuracy,jena_weather/D/short,TiRex,MASE,0.046,,7,4,1
+accuracy,jena_weather/H/long,TiRex,MASE,0.057,,7,5,1
+accuracy,jena_weather/H/medium,TiRex,MASE,0.051,,7,6,1
+accuracy,jena_weather/H/short,TiRex,MASE,0.041,,7,7,1
+accuracy,kdd_cup_2018/D/short,TiRex,MASE,0.381,,7,8,1
+accuracy,kdd_cup_2018/H/long,TiRex,MASE,0.341,,7,9,1
+accuracy,kdd_cup_2018/H/medium,TiRex,MASE,0.337,,7,10,1
+accuracy,kdd_cup_2018/H/short,TiRex,MASE,0.270,,7,11,1
+accuracy,loop_seattle/5T/long,TiRex,MASE,0.090,,7,12,1
+accuracy,loop_seattle/5T/medium,TiRex,MASE,0.083,,7,13,1
+accuracy,loop_seattle/5T/short,TiRex,MASE,0.049,,7,14,1
+accuracy,loop_seattle/D/short,TiRex,MASE,0.042,,7,15,1
+accuracy,loop_seattle/H/long,TiRex,MASE,0.063,,7,16,1
+accuracy,loop_seattle/H/medium,TiRex,MASE,0.065,,7,17,1
+accuracy,loop_seattle/H/short,TiRex,MASE,0.059,,7,18,1
+accuracy,m4_daily/D/short,TiRex,MASE,0.021,,7,19,1
+accuracy,m4_hourly/H/short,TiRex,MASE,0.021,,7,20,1
+accuracy,m4_monthly/M/short,TiRex,MASE,0.093,,7,21,1
+accuracy,m4_quarterly/Q/short,TiRex,MASE,0.074,,7,22,1
+accuracy,m4_weekly/W/short,TiRex,MASE,0.035,,7,23,1
+accuracy,m4_yearly/A/short,TiRex,MASE,0.119,,7,24,1
+accuracy,m_dense/D/short,TiRex,MASE,0.066,,7,25,1
+accuracy,m_dense/H/long,TiRex,MASE,0.122,,7,26,1
+accuracy,m_dense/H/medium,TiRex,MASE,0.121,,7,27,1
+accuracy,m_dense/H/short,TiRex,MASE,0.130,,7,28,1
+accuracy,restaurant/D/short,TiRex,MASE,0.254,,7,29,1
+accuracy,saugeen/D/short,TiRex,MASE,0.382,,7,30,1
+accuracy,saugeen/M/short,TiRex,MASE,0.303,,7,31,1
+accuracy,saugeen/W/short,TiRex,MASE,0.353,,7,32,1
+accuracy,solar/10T/long,TiRex,MASE,0.328,,7,33,1
+accuracy,solar/10T/medium,TiRex,MASE,0.354,,7,34,1
+accuracy,solar/10T/short,TiRex,MASE,0.542,,7,35,1
+accuracy,solar/D/short,TiRex,MASE,0.281,,7,36,1
+accuracy,solar/H/long,TiRex,MASE,0.243,,7,37,1
+accuracy,solar/H/medium,TiRex,MASE,0.260,,7,38,1
+accuracy,solar/H/short,TiRex,MASE,0.259,,7,39,1
+accuracy,solar/W/short,TiRex,MASE,0.154,,7,40,1
+accuracy,sz_taxi/15T/long,TiRex,MASE,0.197,,7,41,1
+accuracy,sz_taxi/15T/medium,TiRex,MASE,0.202,,7,42,1
+accuracy,sz_taxi/15T/short,TiRex,MASE,0.200,,7,43,1
+accuracy,sz_taxi/H/short,TiRex,MASE,0.136,,7,44,1
+accuracy,temperature_rain/D/short,TiRex,MASE,0.550,,7,45,1
+accuracy,us_births/D/short,TiRex,MASE,0.021,,7,46,1
+accuracy,us_births/M/short,TiRex,MASE,0.017,,7,47,1
+accuracy,us_births/W/short,TiRex,MASE,0.013,,7,48,1
diff --git a/result/per_paper/2505.23719/components_architecture.csv b/result/per_paper/2505.23719/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..56cd10e8576070b4e4dbf6905e09f49231b3535e
--- /dev/null
+++ b/result/per_paper/2505.23719/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+TiRex,"A pre-trained time series model based on xLSTM, designed for zero-shot forecasting across short and long horizons.",proposed_here,,"TiRex: We present TiRex, a pre-trained time series model based on xLSTM, which sets a new state of the art in zero-shot forecasting."
+xLSTM,"An enhanced LSTM variant with improved scalability and generalization, incorporating architectural enhancements for in-context learning.",reused_cited,"Beck et al., 2024; Beck et al., 2025","xLSTM has demonstrated in-context learning performance comparable to that of transformer-based large language models (Beck et al., 2025)."
+Contiguous Patch Masking (CPM),A novel training-time masking strategy that enhances xLSTM's state-tracking abilities for long-horizon forecasting.,proposed_here,,CPM enhances xLSTM's ability to produce coherent long-horizon predictions by mitigating degradation common in autoregressive multi-step forecasting.
+Data Augmentation Strategies,"Three augmentation techniques for time series model pre-training, enhancing robustness and performance of TiRex.",proposed_here,,We introduce three augmentation techniques for time series model pre-training and demonstrate their effectiveness in enhancing the robustness and overall performance of TiRex.
diff --git a/result/per_paper/2505.23719/computational.csv b/result/per_paper/2505.23719/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..bac5a541efb26eef38d18e1052982185ab5eeb66
--- /dev/null
+++ b/result/per_paper/2505.23719/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No specific hardware type (e.g., GPU model) is explicitly stated."
+num_devices,,,not_reported,"Number of devices (e.g., GPUs) is not mentioned."
+training_cost,,,not_reported,Monetary or computational cost of training is not reported.
+training_batch_size,,,not_reported,Training batch size is not explicitly stated; tables show inference batch sizes.
+training_steps_or_epochs,,,not_reported,Number of training steps or epochs is not mentioned.
+precision,,,not_reported,"Precision (e.g., FP16/FP32) is not specified."
+inference_latency,,,not_reported,"Latency values depend on batch size (e.g., 0.35 ms per sample at batch size 2^10), but no single value is explicitly stated."
+inference_throughput,,,not_reported,Throughput is not directly reported; inferred from batch size and latency data.
+peak_memory,,,not_reported,"Memory values vary with batch size (e.g., 2.4 GB at batch size 2^10), but peak memory is not explicitly stated."
+flops_or_macs,,,not_reported,FLOPS or MACs are not mentioned.
+num_inference_samples,,,not_reported,Number of inference samples is not specified.
+params,,,not_reported,Model parameter count is not reported.
+context_lengths_evaluated,2048,steps,stated,The text states: 'trained TiRex with a context length of 2048.'
+horizon_lengths_evaluated,32,steps,stated,The text states: 'evaluate samples with a context length of 2048 and a prediction length of 32.'
+inference_batch_size,,,not_reported,"Batch sizes vary (e.g., 2^5 to 2^10), but no specific inference batch size is explicitly stated."
diff --git a/result/per_paper/2506.02081/accuracy_efficiency.csv b/result/per_paper/2506.02081/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f612e5128ed46aa8336f3ba4c17613893fd5f626
--- /dev/null
+++ b/result/per_paper/2506.02081/accuracy_efficiency.csv
@@ -0,0 +1,11 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,ECG,RATFM,accuracy,0.013,
+accuracy,Respiration,RATFM,accuracy,-0.027,
+accuracy,Gait,RATFM,accuracy,0.002,
+accuracy,Satellite,RATFM,accuracy,-0.004,
+accuracy,Power Demand,RATFM,accuracy,-0.011,
+accuracy,Insect EPG,RATFM,accuracy,-0.013,
+accuracy,Atrial BP,RATFM,accuracy,-0.020,
+accuracy,Temperature,RATFM,accuracy,-0.038,
+accuracy,Accelerometer,RATFM,accuracy,-0.034,
+accuracy,Average,RATFM,accuracy,-0.015,
diff --git a/result/per_paper/2506.02081/accuracy_efficiency_traced.csv b/result/per_paper/2506.02081/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..fe61f2789a4d1502b962551903462a6e7fe3a1e7
--- /dev/null
+++ b/result/per_paper/2506.02081/accuracy_efficiency_traced.csv
@@ -0,0 +1,11 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,ECG,RATFM,accuracy,0.013,,6,2,2
+accuracy,Respiration,RATFM,accuracy,-0.027,,6,3,2
+accuracy,Gait,RATFM,accuracy,0.002,,6,4,2
+accuracy,Satellite,RATFM,accuracy,-0.004,,6,5,2
+accuracy,Power Demand,RATFM,accuracy,-0.011,,6,6,2
+accuracy,Insect EPG,RATFM,accuracy,-0.013,,6,7,2
+accuracy,Atrial BP,RATFM,accuracy,-0.020,,6,8,2
+accuracy,Temperature,RATFM,accuracy,-0.038,,6,9,2
+accuracy,Accelerometer,RATFM,accuracy,-0.034,,6,10,2
+accuracy,Average,RATFM,accuracy,-0.015,,6,11,2
diff --git a/result/per_paper/2506.02081/components_architecture.csv b/result/per_paper/2506.02081/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..3f01f62d0800efd381e3b33ca46493b8d1c9001b
--- /dev/null
+++ b/result/per_paper/2506.02081/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Retrieval Module,A component that retrieves relevant examples from a pre-built knowledge base during test-time adaptation to guide the model's anomaly detection process.,proposed_here,,"We propose a retrieval-augmented time series foundation model (RATFM), which enables pretrained time series foundation models to incorporate examples of test-time adaptation."
+Time Series Foundation Model,"A pre-trained model architecture (e.g., Time-MoE or Moment) designed to process and understand time series data, serving as the base for anomaly detection.",reused_cited,"Time-MoE [36], Moment [13]","We conduct experiments on a multi-domain time series dataset, UCR Anomaly Archive [42], using two representative foundation models, Time-MoE [36] and Moment [13]."
+Adaptation Module,A mechanism that integrates retrieved examples into the foundation model's processing pipeline to enable domain-specific adaptation without retraining.,proposed_here,,RATFM enables pretrained time series foundation models to incorporate examples of test-time adaptation.
+Anomaly Detection Head,A task-specific output layer that computes anomaly scores based on the model's internal representations of time series data.,reused_cited,,"Common existing approaches employ Transformer [40] to forecast or reconstruct a time series, identifying anomalies based on the degree of deviation from actual observations [33]."
diff --git a/result/per_paper/2506.02081/computational.csv b/result/per_paper/2506.02081/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e40da2c2719b0bb1c47b018c31b4e2866583faad
--- /dev/null
+++ b/result/per_paper/2506.02081/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware specifications (e.g., GPU/TPU types) are mentioned in the text."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not specified."
+training_cost,,,not_reported,No information about computational or monetary training costs is provided.
+training_batch_size,,,not_reported,Batch sizes for training are not mentioned in the text or tables.
+training_steps_or_epochs,,,not_reported,The number of training steps or epochs is not explicitly reported.
+precision,,,not_reported,"Model precision (e.g., 32-bit vs. 16-bit) is not discussed in the text."
+inference_latency,,,not_reported,Latency metrics for inference are not provided.
+inference_throughput,,,not_reported,Throughput (samples per second) during inference is not mentioned.
+peak_memory,,,not_reported,Peak memory usage during training or inference is not reported.
+flops_or_macs,,,not_reported,FLOPs or MACs (computational operations) are not quantified.
+num_inference_samples,,,not_reported,The number of samples used during inference is not specified.
+params,,,not_reported,"While parameter counts (e.g., 453M for Time-MoE with RATFM training) are listed, the text does not explicitly state the parameter count for the RATFM model itself."
+context_lengths_evaluated,,,not_reported,"Input lengths (e.g., 1120 for Time-MoE) are mentioned, but context lengths specifically evaluated for RATFM are not explicitly stated."
+horizon_lengths_evaluated,96,,stated,"The 'Setting | Lengths' table reports (96, 96) for RATFM (example future time series and forecast target), implying a forecast horizon of 96."
+inference_batch_size,,,not_reported,Inference batch sizes are not mentioned in the text or tables.
diff --git a/result/per_paper/2506.03128/accuracy_efficiency.csv b/result/per_paper/2506.03128/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4dcdd78eb6acbee7bcee1e707023422492918837
--- /dev/null
+++ b/result/per_paper/2506.03128/accuracy_efficiency.csv
@@ -0,0 +1,77 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,
+accuracy,C. Bolt (Small 512),COSMIC (Small),MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC (Small),MASE,512,
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Base),COSMIC (Small),MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Small),MASE,-1.0,
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Large),COSMIC (Small),MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Small),MASE,-1.1,
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,
+accuracy,C. Bolt (Small 512),COSMIC (Small),MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC (Small),MASE,512,
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Base),COSMIC (Small),MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Small),MASE,-1.0,
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Large),COSMIC (Small),MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Small),MASE,-1.1,
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,
+accuracy,C. Bolt (Small 512),COSMIC (Small),MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC (Small),MASE,512,
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Base),COSMIC (Small),MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Small),MASE,-1.0,
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Large),COSMIC (Small),MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Small),MASE,-1.1,
+accuracy,C. Bolt (Small 512),COSMIC (Mini),Agg. Score,512,
+accuracy,C. Bolt (Small 512),COSMIC (Small),Agg. Score,512,
+accuracy,C. Bolt (Base 512),COSMIC (Mini),Agg. Score,512,
+accuracy,C. Bolt (Base 512),COSMIC (Small),Agg. Score,512,
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),Agg. Score,-1.0,
+accuracy,Moirai-1.0 (Base),COSMIC (Small),Agg. Score,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),Agg. Score,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Small),Agg. Score,-1.0,
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),Agg. Score,-1.1,
+accuracy,Moirai-1.1 (Large),COSMIC (Small),Agg. Score,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),Agg. Score,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC (Small),Agg. Score,-1.1,
+accuracy,C. Bolt (Small 512),COSMIC-Mini,MASE,512,
+accuracy,C. Bolt (Small 512),COSMIC-Small,MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC-Mini,MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC-Small,MASE,512,
+accuracy,Moirai-1.0 (Base),COSMIC-Mini,MASE,-1.0,
+accuracy,Moirai-1.0 (Base),COSMIC-Small,MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC-Mini,MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC-Small,MASE,-1.0,
+accuracy,Moirai-1.1 (Large),COSMIC-Mini,MASE,-1.1,
+accuracy,Moirai-1.1 (Large),COSMIC-Small,MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC-Mini,MASE,-1.1,
+accuracy,Moirai-1.1 (Base),COSMIC-Small,MASE,-1.1,
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.0 (Small),COSMIC (Mini),MASE,-1.0,
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,
+accuracy,Moirai-1.1 (Small),COSMIC (Mini),MASE,-1.1,
+accuracy,C. Bolt (Small 512),COSMIC-Mini,MASE,512,
+accuracy,C. Bolt (Base 512),COSMIC-Mini,MASE,512,
+accuracy,Moirai-1.0 (Base),COSMIC-Mini,MASE,-1.0,
+accuracy,Moirai-1.0 (Large),COSMIC-Mini,MASE,-1.0,
+accuracy,Moirai-1.0 (Small),COSMIC-Mini,MASE,-1.0,
+accuracy,Moirai-1.1 (Base),COSMIC-Mini,MASE,-1.1,
+accuracy,Moirai-1.1 (Large),COSMIC-Mini,MASE,-1.1,
+accuracy,Moirai-1.1 (Small),COSMIC-Mini,MASE,-1.1,
diff --git a/result/per_paper/2506.03128/accuracy_efficiency_traced.csv b/result/per_paper/2506.03128/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d16edf2be4b303b8cbd6ce83b11114dc4ac5ce05
--- /dev/null
+++ b/result/per_paper/2506.03128/accuracy_efficiency_traced.csv
@@ -0,0 +1,77 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,,0,7,0
+accuracy,C. Bolt (Small 512),COSMIC (Small),MASE,512,,0,7,0
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,,0,8,0
+accuracy,C. Bolt (Base 512),COSMIC (Small),MASE,512,,0,8,0
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,,0,9,0
+accuracy,Moirai-1.0 (Base),COSMIC (Small),MASE,-1.0,,0,9,0
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,,0,10,0
+accuracy,Moirai-1.0 (Large),COSMIC (Small),MASE,-1.0,,0,10,0
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,,0,11,0
+accuracy,Moirai-1.1 (Large),COSMIC (Small),MASE,-1.1,,0,11,0
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,,0,12,0
+accuracy,Moirai-1.1 (Base),COSMIC (Small),MASE,-1.1,,0,12,0
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,,1,7,0
+accuracy,C. Bolt (Small 512),COSMIC (Small),MASE,512,,1,7,0
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,,1,8,0
+accuracy,C. Bolt (Base 512),COSMIC (Small),MASE,512,,1,8,0
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,,1,9,0
+accuracy,Moirai-1.0 (Base),COSMIC (Small),MASE,-1.0,,1,9,0
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,,1,10,0
+accuracy,Moirai-1.0 (Large),COSMIC (Small),MASE,-1.0,,1,10,0
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,,1,11,0
+accuracy,Moirai-1.1 (Large),COSMIC (Small),MASE,-1.1,,1,11,0
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,,1,12,0
+accuracy,Moirai-1.1 (Base),COSMIC (Small),MASE,-1.1,,1,12,0
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,,2,7,0
+accuracy,C. Bolt (Small 512),COSMIC (Small),MASE,512,,2,7,0
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,,2,8,0
+accuracy,C. Bolt (Base 512),COSMIC (Small),MASE,512,,2,8,0
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,,2,9,0
+accuracy,Moirai-1.0 (Base),COSMIC (Small),MASE,-1.0,,2,9,0
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,,2,10,0
+accuracy,Moirai-1.0 (Large),COSMIC (Small),MASE,-1.0,,2,10,0
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,,2,11,0
+accuracy,Moirai-1.1 (Large),COSMIC (Small),MASE,-1.1,,2,11,0
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,,2,12,0
+accuracy,Moirai-1.1 (Base),COSMIC (Small),MASE,-1.1,,2,12,0
+accuracy,C. Bolt (Small 512),COSMIC (Mini),Agg. Score,512,,4,7,0
+accuracy,C. Bolt (Small 512),COSMIC (Small),Agg. Score,512,,4,7,0
+accuracy,C. Bolt (Base 512),COSMIC (Mini),Agg. Score,512,,4,8,0
+accuracy,C. Bolt (Base 512),COSMIC (Small),Agg. Score,512,,4,8,0
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),Agg. Score,-1.0,,4,9,0
+accuracy,Moirai-1.0 (Base),COSMIC (Small),Agg. Score,-1.0,,4,9,0
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),Agg. Score,-1.0,,4,10,0
+accuracy,Moirai-1.0 (Large),COSMIC (Small),Agg. Score,-1.0,,4,10,0
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),Agg. Score,-1.1,,4,11,0
+accuracy,Moirai-1.1 (Large),COSMIC (Small),Agg. Score,-1.1,,4,11,0
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),Agg. Score,-1.1,,4,12,0
+accuracy,Moirai-1.1 (Base),COSMIC (Small),Agg. Score,-1.1,,4,12,0
+accuracy,C. Bolt (Small 512),COSMIC-Mini,MASE,512,,5,7,0
+accuracy,C. Bolt (Small 512),COSMIC-Small,MASE,512,,5,7,0
+accuracy,C. Bolt (Base 512),COSMIC-Mini,MASE,512,,5,8,0
+accuracy,C. Bolt (Base 512),COSMIC-Small,MASE,512,,5,8,0
+accuracy,Moirai-1.0 (Base),COSMIC-Mini,MASE,-1.0,,5,9,0
+accuracy,Moirai-1.0 (Base),COSMIC-Small,MASE,-1.0,,5,9,0
+accuracy,Moirai-1.0 (Large),COSMIC-Mini,MASE,-1.0,,5,10,0
+accuracy,Moirai-1.0 (Large),COSMIC-Small,MASE,-1.0,,5,10,0
+accuracy,Moirai-1.1 (Large),COSMIC-Mini,MASE,-1.1,,5,11,0
+accuracy,Moirai-1.1 (Large),COSMIC-Small,MASE,-1.1,,5,11,0
+accuracy,Moirai-1.1 (Base),COSMIC-Mini,MASE,-1.1,,5,12,0
+accuracy,Moirai-1.1 (Base),COSMIC-Small,MASE,-1.1,,5,12,0
+accuracy,C. Bolt (Small 512),COSMIC (Mini),MASE,512,,6,8,0
+accuracy,C. Bolt (Base 512),COSMIC (Mini),MASE,512,,6,9,0
+accuracy,Moirai-1.0 (Base),COSMIC (Mini),MASE,-1.0,,6,10,0
+accuracy,Moirai-1.0 (Large),COSMIC (Mini),MASE,-1.0,,6,11,0
+accuracy,Moirai-1.0 (Small),COSMIC (Mini),MASE,-1.0,,6,12,0
+accuracy,Moirai-1.1 (Base),COSMIC (Mini),MASE,-1.1,,6,13,0
+accuracy,Moirai-1.1 (Large),COSMIC (Mini),MASE,-1.1,,6,14,0
+accuracy,Moirai-1.1 (Small),COSMIC (Mini),MASE,-1.1,,6,15,0
+accuracy,C. Bolt (Small 512),COSMIC-Mini,MASE,512,,7,8,0
+accuracy,C. Bolt (Base 512),COSMIC-Mini,MASE,512,,7,9,0
+accuracy,Moirai-1.0 (Base),COSMIC-Mini,MASE,-1.0,,7,10,0
+accuracy,Moirai-1.0 (Large),COSMIC-Mini,MASE,-1.0,,7,11,0
+accuracy,Moirai-1.0 (Small),COSMIC-Mini,MASE,-1.0,,7,12,0
+accuracy,Moirai-1.1 (Base),COSMIC-Mini,MASE,-1.1,,7,13,0
+accuracy,Moirai-1.1 (Large),COSMIC-Mini,MASE,-1.1,,7,14,0
+accuracy,Moirai-1.1 (Small),COSMIC-Mini,MASE,-1.1,,7,15,0
diff --git a/result/per_paper/2506.03128/components_architecture.csv b/result/per_paper/2506.03128/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..851fd51064f691997baf1df64b09894930cab41a
--- /dev/null
+++ b/result/per_paper/2506.03128/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Encoder-Decoder Transformer Architecture,"The core architecture of COSMIC, consisting of an encoder and decoder based on the transformer framework.",reused_cited,"Vaswani et al., 2017","COSMIC utilizes a encoder-decoder transformer architecture (Vaswani et al., 2017)."
+Scaling & Patching,A preprocessing step that normalizes the target time series and covariates individually and divides the time series into patches for input processing.,proposed_here,,"COSMIC normalizes the target time series and covariates individually, often referred to as instance normalization (Kim et al.,"
+Time Encoding,A component that encodes temporal information into the input tokens to capture temporal dependencies.,proposed_here,,"The input layer processes the time series by applying scaling and patching, followed by time and variate encoding to generate input tokens."
+Variate Encoding,A component that encodes covariate information into the input tokens to incorporate auxiliary variables into the model.,proposed_here,,"The input layer processes the time series by applying scaling and patching, followed by time and variate encoding to generate input tokens."
+Output Token Mapping,A post-processing step that maps the forecasted tokens back to the original time series scale and structure to produce the final forecast.,proposed_here,,"The output tokens represent the forecasted patches of the target, and are mapped back to the forecast window."
+Informative Covariate Augmentation,A data augmentation technique that generates synthetic covariates to train COSMIC without requiring real-world datasets containing covariates.,proposed_here,,"To address the challenge of data scarcity, we propose Informative Covariate Augmentation, which enables the training of COSMIC without requiring any datasets that include covariates."
diff --git a/result/per_paper/2506.03128/computational.csv b/result/per_paper/2506.03128/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..3e0383cc745cd7ab94db16851388cf0a35766f58
--- /dev/null
+++ b/result/per_paper/2506.03128/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,No mention of hardware type in the provided text.
+num_devices,,,not_reported,No mention of number of devices in the provided text.
+training_cost,,,not_reported,No mention of training cost in the provided text.
+training_batch_size,,,not_reported,No mention of training batch size in the provided text.
+training_steps_or_epochs,,,not_reported,No mention of training steps or epochs in the provided text.
+precision,,,not_reported,"No mention of precision (e.g., FP16, FP32) in the provided text."
+inference_latency,,,not_reported,No mention of inference latency in the provided text.
+inference_throughput,,,not_reported,No mention of inference throughput in the provided text.
+peak_memory,,,not_reported,No mention of peak memory usage in the provided text.
+flops_or_macs,,,not_reported,No mention of FLOPs or MACs in the provided text.
+num_inference_samples,,,not_reported,No mention of number of inference samples in the provided text.
+params,"20M, 40M, 200M",parameters,stated,"The text states: 'trained three model sizes — mini (20M), small (40M), and base (200M)'"
+context_lengths_evaluated,512,timesteps,stated,The text states: 'A context length of 512 timesteps is used'
+horizon_lengths_evaluated,,,not_reported,"The text lists dataset-specific horizons (e.g., 24, 48) but does not explicitly report a single horizon length for COSMIC."
+inference_batch_size,,,not_reported,No mention of inference batch size in the provided text.
diff --git a/result/per_paper/2506.11029/accuracy_efficiency.csv b/result/per_paper/2506.11029/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6a6eb63866a0562ea11e92a57d934715506526a4
--- /dev/null
+++ b/result/per_paper/2506.11029/accuracy_efficiency.csv
@@ -0,0 +1,95 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,YINGLONG300m,YingLong300m,"MASE, CRPS, Rank",300,
+accuracy,YINGLONG300m,YingLong100m,"MASE, CRPS, Rank",300,
+accuracy,YINGLONG100m,YingLong300m,"MASE, CRPS, Rank",100,
+accuracy,YINGLONG100m,YingLong100m,"MASE, CRPS, Rank",100,
+accuracy,YINGLONG50m,YingLong300m,"MASE, CRPS, Rank",50,
+accuracy,YINGLONG50m,YingLong100m,"MASE, CRPS, Rank",50,
+accuracy,YINGLONG6m,YingLong300m,"MASE, CRPS, Rank",6,
+accuracy,YINGLONG6m,YingLong100m,"MASE, CRPS, Rank",6,
+accuracy,TimesFM-V2500m,YingLong300m,"MASE, CRPS, Rank",2500,
+accuracy,TimesFM-V2500m,YingLong100m,"MASE, CRPS, Rank",2500,
+accuracy,MASE,YingLong300m,MASE,0.658,
+accuracy,MASE,YingLong100m,MASE,0.65,
+accuracy,MASE,YingLong50m,MASE,0.668,
+accuracy,MASE,YingLong60m,MASE,0.709,
+accuracy,CRPS,YingLong300m,MASE,0.421,
+accuracy,CRPS,YingLong100m,MASE,0.422,
+accuracy,CRPS,YingLong50m,MASE,0.43,
+accuracy,CRPS,YingLong60m,MASE,0.459,
+accuracy,Rank,YingLong300m,MASE,5.95,
+accuracy,Rank,YingLong100m,MASE,5.63,
+accuracy,Rank,YingLong50m,MASE,6.19,
+accuracy,Rank,YingLong60m,MASE,9.95,
+accuracy,MASE,YingLong300m,MASE,1.173,
+accuracy,MASE,YingLong100m,MASE,0.695,
+accuracy,MASE,YingLong50m,MASE,0.957,
+accuracy,MASE,YingLong60m,MASE,1.160,
+accuracy,CRPS,YingLong300m,MASE,0.582,
+accuracy,CRPS,YingLong100m,MASE,0.585,
+accuracy,CRPS,YingLong50m,MASE,0.647,
+accuracy,CRPS,YingLong60m,MASE,0.669,
+accuracy,Rank,YingLong300m,MASE,17.67,
+accuracy,Rank,YingLong100m,MASE,19.35,
+accuracy,Rank,YingLong50m,MASE,21.60,
+accuracy,Rank,YingLong60m,MASE,21.28,
+accuracy,MASE,YingLong300m,MASE,0.766,
+accuracy,MASE,YingLong100m,MASE,0.793,
+accuracy,MASE,YingLong50m,MASE,0.799,
+accuracy,MASE,YingLong60m,MASE,0.861,
+accuracy,CRPS,YingLong300m,MASE,0.500,
+accuracy,CRPS,YingLong100m,MASE,0.514,
+accuracy,CRPS,YingLong50m,MASE,0.522,
+accuracy,CRPS,YingLong60m,MASE,0.564,
+accuracy,Rank,YingLong300m,MASE,8.52,
+accuracy,Rank,YingLong100m,MASE,9.96,
+accuracy,Rank,YingLong50m,MASE,10.78,
+accuracy,Rank,YingLong60m,MASE,14.63,
+accuracy,MASE,YingLong300m,MASE,0.829,
+accuracy,MASE,YingLong100m,MASE,0.845,
+accuracy,MASE,YingLong50m,MASE,1.233,
+accuracy,MASE,YingLong60m,MASE,4.000,
+accuracy,CRPS,YingLong300m,MASE,0.569,
+accuracy,CRPS,YingLong100m,MASE,0.683,
+accuracy,CRPS,YingLong50m,MASE,0.831,
+accuracy,CRPS,YingLong60m,MASE,2.464,
+accuracy,Rank,YingLong300m,MASE,13.04,
+accuracy,Rank,YingLong100m,MASE,22.35,
+accuracy,Rank,YingLong50m,MASE,22.93,
+accuracy,Rank,YingLong60m,MASE,21.63,
+accuracy,ETTh1,YingLong6m,MSE,0.415,
+accuracy,ETTh1,YingLong50m,MSE,0.418,
+accuracy,ETTh1,YingLong110m,MSE,0.473,
+accuracy,ETTh1,YingLong300m,MSE,0.444,
+accuracy,ETTh2,YingLong6m,MSE,0.359,
+accuracy,ETTh2,YingLong50m,MSE,0.377,
+accuracy,ETTh2,YingLong110m,MSE,0.392,
+accuracy,ETTh2,YingLong300m,MSE,0.406,
+accuracy,ETTm1,YingLong6m,MSE,0.407,
+accuracy,ETTm1,YingLong50m,MSE,0.385,
+accuracy,ETTm1,YingLong110m,MSE,0.433,
+accuracy,ETTm1,YingLong300m,MSE,0.418,
+accuracy,ETTm2,YingLong6m,MSE,0.303,
+accuracy,ETTm2,YingLong50m,MSE,0.337,
+accuracy,ETTm2,YingLong110m,MSE,0.328,
+accuracy,ETTm2,YingLong300m,MSE,0.346,
+accuracy,Weather,YingLong6m,MSE,0.264,
+accuracy,Weather,YingLong50m,MSE,0.273,
+accuracy,Rank,YingLong6m,MSE,9.25,
+accuracy,Rank,YingLong50m,MSE,6.90,
+accuracy,Rank,YingLong110m,MSE,11.50,
+accuracy,Rank,YingLong300m,MSE,10.50,
+accuracy,YINGLONG $_{300m}$,YingLong-300m,MASE,300,
+accuracy,YINGLONG $_{300m}$,YingLong-110m,MASE,300,
+accuracy,YINGLONG $_{110m}$,YingLong-300m,MASE,110,
+accuracy,YINGLONG $_{110m}$,YingLong-110m,MASE,110,
+accuracy,YINGLONG $_{50m}$,YingLong-300m,MASE,50,
+accuracy,YINGLONG $_{50m}$,YingLong-110m,MASE,50,
+accuracy,TimesFM $_{2500m}$,YingLong-300m,MASE,2500,
+accuracy,TimesFM $_{2500m}$,YingLong-110m,MASE,2500,
+accuracy,YINGLONG $_{6m}$,YingLong-300m,MASE,6,
+accuracy,YINGLONG $_{6m}$,YingLong-110m,MASE,6,
+accuracy,$YINGLONG_{6m}$,YingLong-6m,ETTh1,6,
+accuracy,$YINGLONG_{50m}$,YingLong-6m,ETTh1,50,
+accuracy,$YINGLONG_{100m}$,YingLong-6m,ETTh1,100,
+accuracy,$YINGLONG_{300m}$,YingLong-6m,ETTh1,300,
diff --git a/result/per_paper/2506.11029/accuracy_efficiency_traced.csv b/result/per_paper/2506.11029/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..83a6540a9d6516b83ba82680b2e6688e0f289ba7
--- /dev/null
+++ b/result/per_paper/2506.11029/accuracy_efficiency_traced.csv
@@ -0,0 +1,95 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,YINGLONG300m,YingLong300m,"MASE, CRPS, Rank",300,,1,2,0
+accuracy,YINGLONG300m,YingLong100m,"MASE, CRPS, Rank",300,,1,2,0
+accuracy,YINGLONG100m,YingLong300m,"MASE, CRPS, Rank",100,,1,3,0
+accuracy,YINGLONG100m,YingLong100m,"MASE, CRPS, Rank",100,,1,3,0
+accuracy,YINGLONG50m,YingLong300m,"MASE, CRPS, Rank",50,,1,4,0
+accuracy,YINGLONG50m,YingLong100m,"MASE, CRPS, Rank",50,,1,4,0
+accuracy,YINGLONG6m,YingLong300m,"MASE, CRPS, Rank",6,,1,5,0
+accuracy,YINGLONG6m,YingLong100m,"MASE, CRPS, Rank",6,,1,5,0
+accuracy,TimesFM-V2500m,YingLong300m,"MASE, CRPS, Rank",2500,,1,9,0
+accuracy,TimesFM-V2500m,YingLong100m,"MASE, CRPS, Rank",2500,,1,9,0
+accuracy,MASE,YingLong300m,MASE,0.658,,4,1,1
+accuracy,MASE,YingLong100m,MASE,0.65,,4,1,2
+accuracy,MASE,YingLong50m,MASE,0.668,,4,1,3
+accuracy,MASE,YingLong60m,MASE,0.709,,4,1,4
+accuracy,CRPS,YingLong300m,MASE,0.421,,4,2,1
+accuracy,CRPS,YingLong100m,MASE,0.422,,4,2,2
+accuracy,CRPS,YingLong50m,MASE,0.43,,4,2,3
+accuracy,CRPS,YingLong60m,MASE,0.459,,4,2,4
+accuracy,Rank,YingLong300m,MASE,5.95,,4,3,1
+accuracy,Rank,YingLong100m,MASE,5.63,,4,3,2
+accuracy,Rank,YingLong50m,MASE,6.19,,4,3,3
+accuracy,Rank,YingLong60m,MASE,9.95,,4,3,4
+accuracy,MASE,YingLong300m,MASE,1.173,,4,5,1
+accuracy,MASE,YingLong100m,MASE,0.695,,4,5,2
+accuracy,MASE,YingLong50m,MASE,0.957,,4,5,3
+accuracy,MASE,YingLong60m,MASE,1.160,,4,5,4
+accuracy,CRPS,YingLong300m,MASE,0.582,,4,6,1
+accuracy,CRPS,YingLong100m,MASE,0.585,,4,6,2
+accuracy,CRPS,YingLong50m,MASE,0.647,,4,6,3
+accuracy,CRPS,YingLong60m,MASE,0.669,,4,6,4
+accuracy,Rank,YingLong300m,MASE,17.67,,4,7,1
+accuracy,Rank,YingLong100m,MASE,19.35,,4,7,2
+accuracy,Rank,YingLong50m,MASE,21.60,,4,7,3
+accuracy,Rank,YingLong60m,MASE,21.28,,4,7,4
+accuracy,MASE,YingLong300m,MASE,0.766,,4,9,1
+accuracy,MASE,YingLong100m,MASE,0.793,,4,9,2
+accuracy,MASE,YingLong50m,MASE,0.799,,4,9,3
+accuracy,MASE,YingLong60m,MASE,0.861,,4,9,4
+accuracy,CRPS,YingLong300m,MASE,0.500,,4,10,1
+accuracy,CRPS,YingLong100m,MASE,0.514,,4,10,2
+accuracy,CRPS,YingLong50m,MASE,0.522,,4,10,3
+accuracy,CRPS,YingLong60m,MASE,0.564,,4,10,4
+accuracy,Rank,YingLong300m,MASE,8.52,,4,11,1
+accuracy,Rank,YingLong100m,MASE,9.96,,4,11,2
+accuracy,Rank,YingLong50m,MASE,10.78,,4,11,3
+accuracy,Rank,YingLong60m,MASE,14.63,,4,11,4
+accuracy,MASE,YingLong300m,MASE,0.829,,4,13,1
+accuracy,MASE,YingLong100m,MASE,0.845,,4,13,2
+accuracy,MASE,YingLong50m,MASE,1.233,,4,13,3
+accuracy,MASE,YingLong60m,MASE,4.000,,4,13,4
+accuracy,CRPS,YingLong300m,MASE,0.569,,4,14,1
+accuracy,CRPS,YingLong100m,MASE,0.683,,4,14,2
+accuracy,CRPS,YingLong50m,MASE,0.831,,4,14,3
+accuracy,CRPS,YingLong60m,MASE,2.464,,4,14,4
+accuracy,Rank,YingLong300m,MASE,13.04,,4,15,1
+accuracy,Rank,YingLong100m,MASE,22.35,,4,15,2
+accuracy,Rank,YingLong50m,MASE,22.93,,4,15,3
+accuracy,Rank,YingLong60m,MASE,21.63,,4,15,4
+accuracy,ETTh1,YingLong6m,MSE,0.415,,5,3,9
+accuracy,ETTh1,YingLong50m,MSE,0.418,,5,3,10
+accuracy,ETTh1,YingLong110m,MSE,0.473,,5,3,11
+accuracy,ETTh1,YingLong300m,MSE,0.444,,5,3,12
+accuracy,ETTh2,YingLong6m,MSE,0.359,,5,4,9
+accuracy,ETTh2,YingLong50m,MSE,0.377,,5,4,10
+accuracy,ETTh2,YingLong110m,MSE,0.392,,5,4,11
+accuracy,ETTh2,YingLong300m,MSE,0.406,,5,4,12
+accuracy,ETTm1,YingLong6m,MSE,0.407,,5,5,9
+accuracy,ETTm1,YingLong50m,MSE,0.385,,5,5,10
+accuracy,ETTm1,YingLong110m,MSE,0.433,,5,5,11
+accuracy,ETTm1,YingLong300m,MSE,0.418,,5,5,12
+accuracy,ETTm2,YingLong6m,MSE,0.303,,5,6,9
+accuracy,ETTm2,YingLong50m,MSE,0.337,,5,6,10
+accuracy,ETTm2,YingLong110m,MSE,0.328,,5,6,11
+accuracy,ETTm2,YingLong300m,MSE,0.346,,5,6,12
+accuracy,Weather,YingLong6m,MSE,0.264,,5,7,9
+accuracy,Weather,YingLong50m,MSE,0.273,,5,7,10
+accuracy,Rank,YingLong6m,MSE,9.25,,5,8,9
+accuracy,Rank,YingLong50m,MSE,6.90,,5,8,10
+accuracy,Rank,YingLong110m,MSE,11.50,,5,8,11
+accuracy,Rank,YingLong300m,MSE,10.50,,5,8,12
+accuracy,YINGLONG $_{300m}$,YingLong-300m,MASE,300,,6,2,0
+accuracy,YINGLONG $_{300m}$,YingLong-110m,MASE,300,,6,2,0
+accuracy,YINGLONG $_{110m}$,YingLong-300m,MASE,110,,6,3,0
+accuracy,YINGLONG $_{110m}$,YingLong-110m,MASE,110,,6,3,0
+accuracy,YINGLONG $_{50m}$,YingLong-300m,MASE,50,,6,5,0
+accuracy,YINGLONG $_{50m}$,YingLong-110m,MASE,50,,6,5,0
+accuracy,TimesFM $_{2500m}$,YingLong-300m,MASE,2500,,6,7,0
+accuracy,TimesFM $_{2500m}$,YingLong-110m,MASE,2500,,6,7,0
+accuracy,YINGLONG $_{6m}$,YingLong-300m,MASE,6,,6,11,0
+accuracy,YINGLONG $_{6m}$,YingLong-110m,MASE,6,,6,11,0
+accuracy,$YINGLONG_{6m}$,YingLong-6m,ETTh1,6,,7,3,0
+accuracy,$YINGLONG_{50m}$,YingLong-6m,ETTh1,50,,7,4,0
+accuracy,$YINGLONG_{100m}$,YingLong-6m,ETTh1,100,,7,5,0
+accuracy,$YINGLONG_{300m}$,YingLong-6m,ETTh1,300,,7,6,0
diff --git a/result/per_paper/2506.11029/components_architecture.csv b/result/per_paper/2506.11029/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f9dc33d1e3e0106083f692b5d02a471eb1ac13d6
--- /dev/null
+++ b/result/per_paper/2506.11029/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Transformer Blocks (bidirectional),"Core architecture of YINGLONG, using bidirectional attention for non-causal modeling of time series data.",proposed_here,,"YINGLONG is a non-causal, bidirectional attention encoder-only transformer trained through masked token recovery."
+Joint Forecasting Framework,"A novel forecasting paradigm that integrates historical inputs, target masks, and delayed chain-of-thought (DCoT) tokens for unified prediction.",proposed_here,,Our objective is to integrate the strengths of both recursive and direct forecasting methods by harnessing correlated output modeling while alleviating the constraints of strict causal relationships.
+Delayed Chain of Thought (DCoT),"A mechanism where longer outputs improve accuracy through non-causal, bidirectional reasoning over extended sequences.",proposed_here,,longer outputs significantly enhance model accuracy due to delayed chain-of-thought reasoning in our non-causal approach.
+Multi-input Ensemble,A technique to reduce output variance by combining predictions from multiple input configurations.,proposed_here,,we boost performance by tackling output variance with a multi-input ensemble.
+Non-causal Bidirectional Attention,"Attention mechanism that processes historical and future tokens simultaneously, differing from causal or auto-regressive approaches.",proposed_here,,"YINGLONG is a non-causal, bidirectional attention encoder-only transformer."
+Encoder-only Architecture,"Model structure that uses only encoder layers (no decoder), aligning with language understanding tasks rather than generation.",proposed_here,,aligning more effectively with language understanding tasks than with generation tasks.
diff --git a/result/per_paper/2506.11029/computational.csv b/result/per_paper/2506.11029/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4bb4e56022d4ecf0ea4a58cca24afbd3b61ac788
--- /dev/null
+++ b/result/per_paper/2506.11029/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA A100,,stated,All training is performed on eight NVIDIA A100 at BF16 precision.
+num_devices,8,,stated,All training is performed on eight NVIDIA A100 at BF16 precision.
+training_cost,,,not_reported,
+training_batch_size,512,,stated,The batch size is 512
+training_steps_or_epochs,100000,steps,stated,"We train four models [...] for 100,000 steps."
+precision,BF16,,stated,All training is performed on eight NVIDIA A100 at BF16 precision.
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2506.18046/accuracy_efficiency.csv b/result/per_paper/2506.18046/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..455f7ffe7c1432c10cfe39ff55b5239147739c36
--- /dev/null
+++ b/result/per_paper/2506.18046/accuracy_efficiency.csv
@@ -0,0 +1,67 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,CalIt2,ATrans,accuracy,0.483,
+accuracy,CalIt2,DC,accuracy,0.499,
+accuracy,CalIt2,DLin,accuracy,0.761,
+accuracy,CalIt2,NLin,accuracy,0.694,
+accuracy,CalIt2,Patch,accuracy,0.791,
+accuracy,CalIt2,TsNet,accuracy,0.798,
+accuracy,Non-overlap,ATrans,accuracy,0.527,
+accuracy,Non-overlap,DC,accuracy,0.752,
+accuracy,Non-overlap,DLin,accuracy,0.695,
+accuracy,Non-overlap,NLin,accuracy,0.808,
+accuracy,Non-overlap,Patch,accuracy,0.771,
+accuracy,Daphnet,ATrans,accuracy,0.469,
+accuracy,Daphnet,DC,accuracy,0.486,
+accuracy,Daphnet,DLin,accuracy,0.727,
+accuracy,Daphnet,NLin,accuracy,0.715,
+accuracy,Daphnet,Patch,accuracy,0.739,
+accuracy,Daphnet,TsNet,accuracy,0.773,
+accuracy,Non-overlap,ATrans,accuracy,0.501,
+accuracy,Non-overlap,DC,accuracy,0.728,
+accuracy,Non-overlap,DLin,accuracy,0.715,
+accuracy,Non-overlap,NLin,accuracy,0.741,
+accuracy,Non-overlap,Patch,accuracy,0.754,
+accuracy,MSL,ATrans,accuracy,0.494,
+accuracy,MSL,DC,accuracy,0.502,
+accuracy,MSL,DLin,accuracy,0.626,
+accuracy,MSL,NLin,accuracy,0.594,
+accuracy,MSL,Patch,accuracy,0.637,
+accuracy,MSL,TsNet,accuracy,0.615,
+accuracy,Non-overlap,ATrans,accuracy,0.504,
+accuracy,Non-overlap,DC,accuracy,0.624,
+accuracy,Non-overlap,DLin,accuracy,0.592,
+accuracy,Non-overlap,NLin,accuracy,0.637,
+accuracy,Non-overlap,Patch,accuracy,0.613,
+accuracy,PSM,ATrans,accuracy,0.496,
+accuracy,PSM,DC,accuracy,0.499,
+accuracy,PSM,DLin,accuracy,0.581,
+accuracy,PSM,NLin,accuracy,0.586,
+accuracy,PSM,Patch,accuracy,0.578,
+accuracy,PSM,TsNet,accuracy,0.589,
+accuracy,Non-overlap,ATrans,accuracy,0.499,
+accuracy,Non-overlap,DC,accuracy,0.580,
+accuracy,Non-overlap,DLin,accuracy,0.585,
+accuracy,Non-overlap,NLin,accuracy,0.586,
+accuracy,Non-overlap,Patch,accuracy,0.592,
+accuracy,SKAB,ATrans,accuracy,0.495,
+accuracy,SKAB,DC,accuracy,0.532,
+accuracy,SKAB,DLin,accuracy,0.563,
+accuracy,SKAB,NLin,accuracy,0.558,
+accuracy,SKAB,Patch,accuracy,0.555,
+accuracy,SKAB,TsNet,accuracy,0.592,
+accuracy,Non-overlap,ATrans,accuracy,0.522,
+accuracy,Non-overlap,DC,accuracy,0.593,
+accuracy,Non-overlap,DLin,accuracy,0.583,
+accuracy,Non-overlap,NLin,accuracy,0.597,
+accuracy,Non-overlap,Patch,accuracy,0.620,
+accuracy,SMAP,ATrans,accuracy,0.504,
+accuracy,SMAP,DC,accuracy,0.499,
+accuracy,SMAP,DLin,accuracy,0.398,
+accuracy,SMAP,NLin,accuracy,0.434,
+accuracy,SMAP,Patch,accuracy,0.449,
+accuracy,SMAP,TsNet,accuracy,0.455,
+accuracy,Non-overlap,ATrans,accuracy,0.516,
+accuracy,Non-overlap,DC,accuracy,0.397,
+accuracy,Non-overlap,DLin,accuracy,0.434,
+accuracy,Non-overlap,NLin,accuracy,0.448,
+accuracy,Non-overlap,Patch,accuracy,0.453,
diff --git a/result/per_paper/2506.18046/accuracy_efficiency_traced.csv b/result/per_paper/2506.18046/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b5a185d9651b341b3e7b210f121c9fa9dbf0aadf
--- /dev/null
+++ b/result/per_paper/2506.18046/accuracy_efficiency_traced.csv
@@ -0,0 +1,67 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,CalIt2,ATrans,accuracy,0.483,,1,1,2
+accuracy,CalIt2,DC,accuracy,0.499,,1,1,3
+accuracy,CalIt2,DLin,accuracy,0.761,,1,1,4
+accuracy,CalIt2,NLin,accuracy,0.694,,1,1,5
+accuracy,CalIt2,Patch,accuracy,0.791,,1,1,6
+accuracy,CalIt2,TsNet,accuracy,0.798,,1,1,7
+accuracy,Non-overlap,ATrans,accuracy,0.527,,1,2,2
+accuracy,Non-overlap,DC,accuracy,0.752,,1,2,3
+accuracy,Non-overlap,DLin,accuracy,0.695,,1,2,4
+accuracy,Non-overlap,NLin,accuracy,0.808,,1,2,5
+accuracy,Non-overlap,Patch,accuracy,0.771,,1,2,6
+accuracy,Daphnet,ATrans,accuracy,0.469,,1,3,2
+accuracy,Daphnet,DC,accuracy,0.486,,1,3,3
+accuracy,Daphnet,DLin,accuracy,0.727,,1,3,4
+accuracy,Daphnet,NLin,accuracy,0.715,,1,3,5
+accuracy,Daphnet,Patch,accuracy,0.739,,1,3,6
+accuracy,Daphnet,TsNet,accuracy,0.773,,1,3,7
+accuracy,Non-overlap,ATrans,accuracy,0.501,,1,4,2
+accuracy,Non-overlap,DC,accuracy,0.728,,1,4,3
+accuracy,Non-overlap,DLin,accuracy,0.715,,1,4,4
+accuracy,Non-overlap,NLin,accuracy,0.741,,1,4,5
+accuracy,Non-overlap,Patch,accuracy,0.754,,1,4,6
+accuracy,MSL,ATrans,accuracy,0.494,,1,5,2
+accuracy,MSL,DC,accuracy,0.502,,1,5,3
+accuracy,MSL,DLin,accuracy,0.626,,1,5,4
+accuracy,MSL,NLin,accuracy,0.594,,1,5,5
+accuracy,MSL,Patch,accuracy,0.637,,1,5,6
+accuracy,MSL,TsNet,accuracy,0.615,,1,5,7
+accuracy,Non-overlap,ATrans,accuracy,0.504,,1,6,2
+accuracy,Non-overlap,DC,accuracy,0.624,,1,6,3
+accuracy,Non-overlap,DLin,accuracy,0.592,,1,6,4
+accuracy,Non-overlap,NLin,accuracy,0.637,,1,6,5
+accuracy,Non-overlap,Patch,accuracy,0.613,,1,6,6
+accuracy,PSM,ATrans,accuracy,0.496,,1,7,2
+accuracy,PSM,DC,accuracy,0.499,,1,7,3
+accuracy,PSM,DLin,accuracy,0.581,,1,7,4
+accuracy,PSM,NLin,accuracy,0.586,,1,7,5
+accuracy,PSM,Patch,accuracy,0.578,,1,7,6
+accuracy,PSM,TsNet,accuracy,0.589,,1,7,7
+accuracy,Non-overlap,ATrans,accuracy,0.499,,1,8,2
+accuracy,Non-overlap,DC,accuracy,0.580,,1,8,3
+accuracy,Non-overlap,DLin,accuracy,0.585,,1,8,4
+accuracy,Non-overlap,NLin,accuracy,0.586,,1,8,5
+accuracy,Non-overlap,Patch,accuracy,0.592,,1,8,6
+accuracy,SKAB,ATrans,accuracy,0.495,,1,9,2
+accuracy,SKAB,DC,accuracy,0.532,,1,9,3
+accuracy,SKAB,DLin,accuracy,0.563,,1,9,4
+accuracy,SKAB,NLin,accuracy,0.558,,1,9,5
+accuracy,SKAB,Patch,accuracy,0.555,,1,9,6
+accuracy,SKAB,TsNet,accuracy,0.592,,1,9,7
+accuracy,Non-overlap,ATrans,accuracy,0.522,,1,10,2
+accuracy,Non-overlap,DC,accuracy,0.593,,1,10,3
+accuracy,Non-overlap,DLin,accuracy,0.583,,1,10,4
+accuracy,Non-overlap,NLin,accuracy,0.597,,1,10,5
+accuracy,Non-overlap,Patch,accuracy,0.620,,1,10,6
+accuracy,SMAP,ATrans,accuracy,0.504,,1,11,2
+accuracy,SMAP,DC,accuracy,0.499,,1,11,3
+accuracy,SMAP,DLin,accuracy,0.398,,1,11,4
+accuracy,SMAP,NLin,accuracy,0.434,,1,11,5
+accuracy,SMAP,Patch,accuracy,0.449,,1,11,6
+accuracy,SMAP,TsNet,accuracy,0.455,,1,11,7
+accuracy,Non-overlap,ATrans,accuracy,0.516,,1,12,2
+accuracy,Non-overlap,DC,accuracy,0.397,,1,12,3
+accuracy,Non-overlap,DLin,accuracy,0.434,,1,12,4
+accuracy,Non-overlap,NLin,accuracy,0.448,,1,12,5
+accuracy,Non-overlap,Patch,accuracy,0.453,,1,12,6
diff --git a/result/per_paper/2506.18046/components_architecture.csv b/result/per_paper/2506.18046/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a8fe1d17f490c392d54a6ea8416b65d0b0a40df3
--- /dev/null
+++ b/result/per_paper/2506.18046/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Benchmark Dataset Collection,"A collection of 29 public multivariate datasets and 1,635 univariate time series from diverse domains to enable comprehensive evaluations.",proposed_here,"Qiu et al., 2025","TAB encompasses 29 public multivariate datasets and 1,635 univariate time series from different domains to facilitate more comprehensive evaluations on diverse datasets."
+TSAD Methods Coverage,"A categorization of TSAD methods including Non-learning, Machine learning, Deep learning, LLM-based, and Time-series pre-trained methods.",proposed_here,"Qiu et al., 2025","TAB covers a variety of TSAD methods, including Non-learning, Machine learning, Deep learning, LLM-based, and Time-series pre-trained methods."
+Unified Evaluation Pipeline,An automated and standardized evaluation framework for fair and easy comparison of TSAD methods.,proposed_here,"Qiu et al., 2025",TAB features a unified and automated evaluation pipeline that enables fair and easy evaluation of TSAD methods.
+Evaluation Results Reporting,A systematic analysis and reporting of performance outcomes of existing TSAD methods using the TAB benchmark.,proposed_here,"Qiu et al., 2025","Finally, we employ TAB to evaluate existing TSAD methods and report on the outcomes, thereby offering a deeper insight into the performance of these methods."
diff --git a/result/per_paper/2506.18046/computational.csv b/result/per_paper/2506.18046/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e42a22f1e724efb36da9ee16a6d20710701025c1
--- /dev/null
+++ b/result/per_paper/2506.18046/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware specifications (e.g., GPU/TPU types) are mentioned for the TAB model."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not reported."
+training_cost,,,not_reported,"Training cost (e.g., computational resources, monetary cost) is not mentioned."
+training_batch_size,,,not_reported,Batch size during training is not specified.
+training_steps_or_epochs,,,not_reported,Training duration (steps/epochs) is not reported.
+precision,,,not_reported,"Precision metrics are discussed in the evaluation context, but not as computational parameters for the TAB model itself."
+inference_latency,,,not_reported,"Inference latency (e.g., time per sample) is not mentioned."
+inference_throughput,,,not_reported,"Inference throughput (e.g., samples/second) is not reported."
+peak_memory,,,not_reported,Peak memory usage during training or inference is not specified.
+flops_or_macs,,,not_reported,FLOPs or MACs (computational complexity) are not mentioned.
+num_inference_samples,,,not_reported,Number of samples used for inference evaluation is not reported.
+params,,,not_reported,Number of parameters in the TAB model is not specified.
+context_lengths_evaluated,,,not_reported,"Context lengths (e.g., sequence lengths) evaluated are not mentioned."
+horizon_lengths_evaluated,,,not_reported,"Horizon lengths (e.g., prediction window) evaluated are not reported."
+inference_batch_size,,,not_reported,Batch size during inference is not specified.
diff --git a/result/per_paper/2506.23596/accuracy_efficiency.csv b/result/per_paper/2506.23596/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ac4033644877981f19752ad3e6bf7c9426750b2d
--- /dev/null
+++ b/result/per_paper/2506.23596/accuracy_efficiency.csv
@@ -0,0 +1,40 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,DC,Anomaly-to-Prompt,F1,59.59,3.94
+accuracy,CAD,Anomaly-to-Prompt,F1,53.75,0.11
+accuracy,DC,Anomaly-to-Prompt,F1,57.90,1.95
+accuracy,CAD,Anomaly-to-Prompt,F1,57.42,0.52
+accuracy,DC,Anomaly-to-Prompt,F1,55.13,3.80
+accuracy,CAD,Anomaly-to-Prompt,F1,16.64,8.75
+accuracy,DC,Anomaly-to-Prompt,F1,53.07,5.30
+accuracy,CAD,Anomaly-to-Prompt,F1,39.76,6.38
+accuracy,DC,Anomaly-to-Prompt,F1,61.39,4.89
+accuracy,CAD,Anomaly-to-Prompt,F1,23.72,9.43
+accuracy,A2P (Ours),Anomaly-to-Prompt,F1,67.55,5.62
+accuracy,DC,Anomaly-to-Prompt,F1,59.04,3.69
+accuracy,CAD,Anomaly-to-Prompt,F1,16.87,5.21
+accuracy,DC,Anomaly-to-Prompt,F1,57.80,1.32
+accuracy,CAD,Anomaly-to-Prompt,F1,50.46,0.82
+accuracy,DC,Anomaly-to-Prompt,F1,55.44,2.84
+accuracy,CAD,Anomaly-to-Prompt,F1,26.24,1.78
+accuracy,DC,Anomaly-to-Prompt,F1,52.53,7.47
+accuracy,CAD,Anomaly-to-Prompt,F1,21.14,7.35
+accuracy,DC,Anomaly-to-Prompt,F1,51.21,7.64
+accuracy,CAD,Anomaly-to-Prompt,F1,51.44,14.45
+accuracy,A2P (Ours),Anomaly-to-Prompt,F1,74.63,5.92
+accuracy,DC,Anomaly-to-Prompt,F1,56.89,8.21
+accuracy,CAD,Anomaly-to-Prompt,F1,20.04,9.04
+accuracy,DC,Anomaly-to-Prompt,F1,56.08,5.30
+accuracy,CAD,Anomaly-to-Prompt,F1,21.12,1.87
+accuracy,DC,Anomaly-to-Prompt,F1,57.70,10.28
+accuracy,CAD,Anomaly-to-Prompt,F1,3.57,2.25
+accuracy,DC,Anomaly-to-Prompt,F1,49.72,4.23
+accuracy,CAD,Anomaly-to-Prompt,F1,24.87,6.02
+accuracy,DC,Anomaly-to-Prompt,F1,54.32,11.29
+accuracy,CAD,Anomaly-to-Prompt,F1,66.33,4.21
+accuracy,A2P (Ours),Anomaly-to-Prompt,F1,69.35,7.15
+accuracy,GPT2,P-TST,F1,2,
+accuracy,GPT2,DC,F1,2,
+accuracy,A2P (Ours),P-TST,F1,2,
+accuracy,A2P (Ours),DC,F1,2,
+accuracy,GPT2,A2P,MSE,2,
+accuracy,A2P (Ours),A2P,MSE,2,
diff --git a/result/per_paper/2506.23596/accuracy_efficiency_traced.csv b/result/per_paper/2506.23596/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..819a0640abd9112e018f23cc10d7b83ee74334e1
--- /dev/null
+++ b/result/per_paper/2506.23596/accuracy_efficiency_traced.csv
@@ -0,0 +1,40 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,DC,Anomaly-to-Prompt,F1,59.59,3.94,0,3,1
+accuracy,CAD,Anomaly-to-Prompt,F1,53.75,0.11,0,4,1
+accuracy,DC,Anomaly-to-Prompt,F1,57.90,1.95,0,6,1
+accuracy,CAD,Anomaly-to-Prompt,F1,57.42,0.52,0,7,1
+accuracy,DC,Anomaly-to-Prompt,F1,55.13,3.80,0,9,1
+accuracy,CAD,Anomaly-to-Prompt,F1,16.64,8.75,0,10,1
+accuracy,DC,Anomaly-to-Prompt,F1,53.07,5.30,0,12,1
+accuracy,CAD,Anomaly-to-Prompt,F1,39.76,6.38,0,13,1
+accuracy,DC,Anomaly-to-Prompt,F1,61.39,4.89,0,15,1
+accuracy,CAD,Anomaly-to-Prompt,F1,23.72,9.43,0,16,1
+accuracy,A2P (Ours),Anomaly-to-Prompt,F1,67.55,5.62,0,17,1
+accuracy,DC,Anomaly-to-Prompt,F1,59.04,3.69,0,19,1
+accuracy,CAD,Anomaly-to-Prompt,F1,16.87,5.21,0,20,1
+accuracy,DC,Anomaly-to-Prompt,F1,57.80,1.32,0,22,1
+accuracy,CAD,Anomaly-to-Prompt,F1,50.46,0.82,0,23,1
+accuracy,DC,Anomaly-to-Prompt,F1,55.44,2.84,0,25,1
+accuracy,CAD,Anomaly-to-Prompt,F1,26.24,1.78,0,26,1
+accuracy,DC,Anomaly-to-Prompt,F1,52.53,7.47,0,28,1
+accuracy,CAD,Anomaly-to-Prompt,F1,21.14,7.35,0,29,1
+accuracy,DC,Anomaly-to-Prompt,F1,51.21,7.64,0,31,1
+accuracy,CAD,Anomaly-to-Prompt,F1,51.44,14.45,0,32,1
+accuracy,A2P (Ours),Anomaly-to-Prompt,F1,74.63,5.92,0,33,1
+accuracy,DC,Anomaly-to-Prompt,F1,56.89,8.21,0,35,1
+accuracy,CAD,Anomaly-to-Prompt,F1,20.04,9.04,0,36,1
+accuracy,DC,Anomaly-to-Prompt,F1,56.08,5.30,0,38,1
+accuracy,CAD,Anomaly-to-Prompt,F1,21.12,1.87,0,39,1
+accuracy,DC,Anomaly-to-Prompt,F1,57.70,10.28,0,41,1
+accuracy,CAD,Anomaly-to-Prompt,F1,3.57,2.25,0,42,1
+accuracy,DC,Anomaly-to-Prompt,F1,49.72,4.23,0,44,1
+accuracy,CAD,Anomaly-to-Prompt,F1,24.87,6.02,0,45,1
+accuracy,DC,Anomaly-to-Prompt,F1,54.32,11.29,0,47,1
+accuracy,CAD,Anomaly-to-Prompt,F1,66.33,4.21,0,48,1
+accuracy,A2P (Ours),Anomaly-to-Prompt,F1,69.35,7.15,0,49,1
+accuracy,GPT2,P-TST,F1,2,,1,8,0
+accuracy,GPT2,DC,F1,2,,1,8,0
+accuracy,A2P (Ours),P-TST,F1,2,,1,17,0
+accuracy,A2P (Ours),DC,F1,2,,1,17,0
+accuracy,GPT2,A2P,MSE,2,,6,3,0
+accuracy,A2P (Ours),A2P,MSE,2,,6,6,0
diff --git a/result/per_paper/2506.23596/components_architecture.csv b/result/per_paper/2506.23596/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..57bfbaf3467157055ca6dccee4b0532bf17a80c0
--- /dev/null
+++ b/result/per_paper/2506.23596/components_architecture.csv
@@ -0,0 +1,4 @@
+component,what_it_is,provenance,citation,evidence
+Anomaly-Aware Forecasting (AAF),A forecasting component that incorporates anomaly awareness during training to predict future time points with anomalies by learning relationships between anomalies and normal signals.,proposed_here,,"To enable the forecasting model to forecast abnormal time points, we adopt a strategy to learn the relationships of anomalies."
+Synthetic Anomaly Prompting (SAP),A module that introduces a learnable Anomaly Prompt Pool (APP) to simulate diverse anomaly patterns using signal-adaptive prompts for robust anomaly detection.,proposed_here,,"For the robust detection of anomalies, our proposed SAP introduces a learnable Anomaly Prompt Pool (APP) that simulates diverse anomaly patterns using signal-adaptive prompt."
+Anomaly Prompt Pool (APP),A learnable component within SAP that stores and generates synthetic anomaly patterns to guide the model in detecting future anomalies.,proposed_here,,our proposed SAP introduces a learnable Anomaly Prompt Pool (APP) that simulates diverse anomaly patterns using signal-adaptive prompt.
diff --git a/result/per_paper/2506.23596/computational.csv b/result/per_paper/2506.23596/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..9aa25bc1a76bd9e0912031b8c79deaf29ae0da49
--- /dev/null
+++ b/result/per_paper/2506.23596/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA RTX 3090,GPU,stated,experiments were executed on single GPU (NVIDIA RTX 3090)
+num_devices,1,device,stated,experiments were executed on single GPU (NVIDIA RTX 3090)
+training_cost,,,not_reported,
+training_batch_size,16,samples,stated,batch size of 16
+training_steps_or_epochs,5,epochs,stated,trained the models for 5 epochs
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,100,timesteps,stated,input sequence length $L_{in}$ as 100
+horizon_lengths_evaluated,100-400,timesteps,stated,output sequence length $L_{out}$ from 100 to 400
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2507.01875/accuracy_efficiency.csv b/result/per_paper/2507.01875/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2507.01875/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2507.01875/accuracy_efficiency_traced.csv b/result/per_paper/2507.01875/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2507.01875/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2507.01875/components_architecture.csv b/result/per_paper/2507.01875/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4d66524d348069095bab1ea75f99b92f78892cbd
--- /dev/null
+++ b/result/per_paper/2507.01875/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Encoder,A causal dilated convolutional network that processes input time-series data into latent representations.,proposed_here,,"Figure 1: FAE's encoder/decoder architecture using causal dilated convolutions, implemented through a stack of 1D convolutional layers."
+Decoder,A dilated convolutional network that reconstructs time-series data from latent representations generated by the encoder.,proposed_here,,"Figure 1: FAE's encoder/decoder architecture using causal dilated convolutions, implemented through a stack of 1D convolutional layers."
+Causal Dilated Convolutional Layers,"A stack of 1D convolutional layers with causal dilation, enabling the model to capture long-range temporal dependencies while preserving temporal order.",reused_cited,"Gastón García González et al., 2023 [13]","The encoder/decoder architecture uses causal dilated convolutions, implemented through a stack of 1D convolutional layers."
+Variational Auto-Encoder (VAE),A probabilistic generative model that learns latent representations of time-series data by maximizing the evidence lower bound (ELBO).,reused_cited,"Kingma & Welling, 2013 [23]",FAE is based on Variational Auto-Encoders (VAEs).
+Dilated Convolutional Neural Network (DCNN),A neural network architecture with dilated convolutions that expands the receptive field without increasing the number of parameters.,reused_cited,"Gastón García González et al., 2023 [13]",FAE leverages VAEs and Dilated Convolutional Neural Networks (DCNNs) to build a generic model.
diff --git a/result/per_paper/2507.01875/computational.csv b/result/per_paper/2507.01875/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..906776378bb328078bdfd789fd35c66fd8c64f44
--- /dev/null
+++ b/result/per_paper/2507.01875/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware type (e.g., GPU/TPU) is explicitly mentioned."
+num_devices,,,not_reported,"No information about the number of devices (e.g., GPUs) is provided."
+training_cost,,,not_reported,"No training cost (e.g., monetary or computational budget) is mentioned."
+training_batch_size,32,samples,stated,Table 1 lists 'm - mini-batch size' with best value 32.
+training_steps_or_epochs,,,not_reported,No training steps or epochs are explicitly reported.
+precision,,,not_reported,"No precision (e.g., 32-bit/16-bit) is mentioned."
+inference_latency,,,not_reported,No inference latency is reported.
+inference_throughput,,,not_reported,No inference throughput is reported.
+peak_memory,,,not_reported,No peak memory usage is mentioned.
+flops_or_macs,,,not_reported,No FLOPs or MACs are reported.
+num_inference_samples,,,not_reported,No number of inference samples is specified.
+params,483840,parameters,stated,Text states 'the exact number of trainable parameters in the identified architecture is p = 483.840'.
+context_lengths_evaluated,256,samples,stated,Table 1 lists 'T - sequence length' with best value 256.
+horizon_lengths_evaluated,,,not_reported,No explicit horizon length (forecasting window) is reported.
+inference_batch_size,,,not_reported,No inference batch size is mentioned.
diff --git a/result/per_paper/2508.02879/accuracy_efficiency.csv b/result/per_paper/2508.02879/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b65cd66602c9b992ca49736724760007365a74f5
--- /dev/null
+++ b/result/per_paper/2508.02879/accuracy_efficiency.csv
@@ -0,0 +1,32 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Device,CauKer100K,accuracy,0.7209,
+accuracy,ECG,CauKer100K,accuracy,0.8539,
+accuracy,EOG,CauKer100K,accuracy,0.5000,
+accuracy,EPG,CauKer100K,accuracy,0.9980,
+accuracy,HRM,CauKer100K,accuracy,0.8602,
+accuracy,Hemodynamics,CauKer100K,accuracy,0.7131,
+accuracy,Image,CauKer100K,accuracy,0.7798,
+accuracy,Motion,CauKer100K,accuracy,0.7883,
+accuracy,Power,CauKer100K,accuracy,0.9667,
+accuracy,Sensor,CauKer100K,accuracy,0.7823,
+accuracy,Simulated,CauKer100K,accuracy,0.9434,
+accuracy,Spectro,CauKer100K,accuracy,0.7226,
+accuracy,Spectrum,CauKer100K,accuracy,0.8126,
+accuracy,Traffic,CauKer100K,accuracy,0.9120,
+accuracy,Trajectory,CauKer100K,accuracy,0.5385,
+accuracy,MOMENT (77M),CauKer,UCR Accuracy (%),77,
+accuracy,Mantis (8M),CauKer,UCR Accuracy (%),8,
+accuracy,CAP (EEG),CauKer100K,Accuracy,0.782,
+accuracy,HAR,CauKer100K,Accuracy,0.946,
+accuracy,MI (EEG),CauKer100K,Accuracy,0.563,
+accuracy,SEDFx (EEG),CauKer100K,Accuracy,0.7700,
+accuracy,Win counts (out of 17; ties counted),CauKer100K,Accuracy,11,
+accuracy,Average over all 17 datasets,CauKer100K,Accuracy,0.820,
+accuracy,CauKer1M,CauKer1M,MASE,1,
+accuracy,CauKer1M,CauKer1M,MASE,1,
+accuracy,CauKer1M,CauKer1M,MASE,1,
+accuracy,CauKer1M,CauKer1M,MASE,1,
+accuracy,Mantis (CAUKER-100K),CauKer-100K,AUROC,0.7984,
+accuracy,Mantis (CAUKER-1M),CauKer-100K,AUROC,0.8189,
+accuracy,Mantis (CAUKER-100K),CauKer-100K,AUROC,0.8534,
+accuracy,Mantis (CAUKER-1M),CauKer-100K,AUROC,0.8709,
diff --git a/result/per_paper/2508.02879/accuracy_efficiency_traced.csv b/result/per_paper/2508.02879/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..95100e2c51332978b3cbb76c027bf44438596984
--- /dev/null
+++ b/result/per_paper/2508.02879/accuracy_efficiency_traced.csv
@@ -0,0 +1,32 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Device,CauKer100K,accuracy,0.7209,,1,1,1
+accuracy,ECG,CauKer100K,accuracy,0.8539,,1,2,1
+accuracy,EOG,CauKer100K,accuracy,0.5000,,1,3,1
+accuracy,EPG,CauKer100K,accuracy,0.9980,,1,4,1
+accuracy,HRM,CauKer100K,accuracy,0.8602,,1,5,1
+accuracy,Hemodynamics,CauKer100K,accuracy,0.7131,,1,6,1
+accuracy,Image,CauKer100K,accuracy,0.7798,,1,7,1
+accuracy,Motion,CauKer100K,accuracy,0.7883,,1,8,1
+accuracy,Power,CauKer100K,accuracy,0.9667,,1,9,1
+accuracy,Sensor,CauKer100K,accuracy,0.7823,,1,10,1
+accuracy,Simulated,CauKer100K,accuracy,0.9434,,1,11,1
+accuracy,Spectro,CauKer100K,accuracy,0.7226,,1,12,1
+accuracy,Spectrum,CauKer100K,accuracy,0.8126,,1,13,1
+accuracy,Traffic,CauKer100K,accuracy,0.9120,,1,14,1
+accuracy,Trajectory,CauKer100K,accuracy,0.5385,,1,15,1
+accuracy,MOMENT (77M),CauKer,UCR Accuracy (%),77,,2,1,0
+accuracy,Mantis (8M),CauKer,UCR Accuracy (%),8,,2,11,0
+accuracy,CAP (EEG),CauKer100K,Accuracy,0.782,,5,1,2
+accuracy,HAR,CauKer100K,Accuracy,0.946,,5,2,2
+accuracy,MI (EEG),CauKer100K,Accuracy,0.563,,5,3,2
+accuracy,SEDFx (EEG),CauKer100K,Accuracy,0.7700,,5,4,2
+accuracy,Win counts (out of 17; ties counted),CauKer100K,Accuracy,11,,5,5,2
+accuracy,Average over all 17 datasets,CauKer100K,Accuracy,0.820,,5,6,2
+accuracy,CauKer1M,CauKer1M,MASE,1,,6,2,0
+accuracy,CauKer1M,CauKer1M,MASE,1,,6,4,0
+accuracy,CauKer1M,CauKer1M,MASE,1,,6,6,0
+accuracy,CauKer1M,CauKer1M,MASE,1,,6,8,0
+accuracy,Mantis (CAUKER-100K),CauKer-100K,AUROC,0.7984,,7,2,1
+accuracy,Mantis (CAUKER-1M),CauKer-100K,AUROC,0.8189,,7,3,1
+accuracy,Mantis (CAUKER-100K),CauKer-100K,AUROC,0.8534,,7,5,1
+accuracy,Mantis (CAUKER-1M),CauKer-100K,AUROC,0.8709,,7,6,1
diff --git a/result/per_paper/2508.02879/components_architecture.csv b/result/per_paper/2508.02879/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..09f777c42a9a195b65d9b90f25d7b44b48f7fcaf
--- /dev/null
+++ b/result/per_paper/2508.02879/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Gaussian Process (GP) Kernel Composition,A method for generating synthetic time series with realistic trends and nonlinear interactions by combining GP kernels.,reused_cited,"Rasmussen, C. E., & Williams, C. K. I. (2006). Gaussian Processes for Machine Learning.",CAUKER combines Gaussian Process (GP) kernel composition with Structural Causal Models (SCM) to produce data for sample-efficient pre-training.
+Structural Causal Models (SCM),A framework for modeling causal relationships and generating synthetic time series with seasonality and nonlinear interactions.,reused_cited,"Pearl, J. (2009). Causality: Models, Reasoning, and Inference.",CAUKER combines Gaussian Process (GP) kernel composition with Structural Causal Models (SCM) to produce data for sample-efficient pre-training.
+Synthetic Data Generation Pipeline,"A novel algorithm for generating diverse, causally coherent synthetic time series with realistic temporal dependencies.",proposed_here,,"We propose CAUKER, a novel algorithm designed to generate diverse, causally coherent synthetic time series with realistic trends, seasonality, and nonlinear interactions."
+Time Series Foundation Model (TSFM) Pre-Training Framework,"A framework for pre-training classification TSFMs using synthetic data generated by CAUKER, enabling sample-efficient training.",proposed_here,,Our experiments reveal that CAUKER-generated datasets exhibit clear scaling laws for both dataset size and model capacity.
diff --git a/result/per_paper/2508.02879/computational.csv b/result/per_paper/2508.02879/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5b8b786d8a2f68acfdb01fa74e2c4af2c0bd46a6
--- /dev/null
+++ b/result/per_paper/2508.02879/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No specific hardware type (e.g., GPU/CPU) is mentioned in the text or tables."
+num_devices,,,not_reported,"No information about the number of devices (e.g., GPUs) used is provided."
+training_cost,,,not_reported,"Training cost (e.g., monetary or computational resources) is not quantified in the text or tables."
+training_batch_size,,,not_reported,Batch size during training is not mentioned.
+training_steps_or_epochs,,,not_reported,Number of training steps or epochs is not reported.
+precision,,,not_reported,"Precision (e.g., FP16, FP32) is not specified."
+inference_latency,,,not_reported,Inference latency is not mentioned.
+inference_throughput,,,not_reported,Inference throughput is not quantified.
+peak_memory,,,not_reported,Peak memory usage is not reported.
+flops_or_macs,,,not_reported,FLOPs or MACs are not mentioned.
+num_inference_samples,,,not_reported,Number of inference samples is not specified.
+params,,,not_reported,Model parameter count is not reported.
+context_lengths_evaluated,,,not_reported,Context lengths evaluated are not mentioned.
+horizon_lengths_evaluated,,,not_reported,Horizon lengths evaluated are not specified.
+inference_batch_size,,,not_reported,Inference batch size is not reported.
diff --git a/result/per_paper/2508.04379/accuracy_efficiency.csv b/result/per_paper/2508.04379/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ff241b1b6dad29d1789226347e9a8e5d50b823d0
--- /dev/null
+++ b/result/per_paper/2508.04379/accuracy_efficiency.csv
@@ -0,0 +1,132 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,M1 Monthly,VisionTS++,MAE,1919.97,
+accuracy,M1 Monthly,VisionTS++,MAE,1,
+accuracy,M3 Monthly,VisionTS++,MAE,591.44,
+accuracy,M3 Monthly,VisionTS++,MAE,581.68,
+accuracy,M3 Other,VisionTS++,MAE,180.99,
+accuracy,M3 Other,VisionTS++,MAE,186.13,
+accuracy,M4 Monthly,VisionTS++,MAE,533.16,
+accuracy,M4 Monthly,VisionTS++,MAE,533.74,
+accuracy,M4 Weekly,VisionTS++,MAE,281.76,
+accuracy,M4 Weekly,VisionTS++,MAE,280.88,
+accuracy,M4 Daily,VisionTS++,MAE,190.54,
+accuracy,M4 Daily,VisionTS++,MAE,172.31,
+accuracy,M4 Hourly,VisionTS++,MAE,169.17,
+accuracy,M4 Hourly,VisionTS++,MAE,202.99,
+accuracy,Tourism Quarterly,VisionTS++,MAE,5623.23,
+accuracy,Tourism Quarterly,VisionTS++,MAE,6,
+accuracy,Tourism Monthly,VisionTS++,MAE,1667.65,
+accuracy,Tourism Monthly,VisionTS++,MAE,2,
+accuracy,CIF 2016,VisionTS++,MAE,5664488.37,
+accuracy,CIF 2016,VisionTS++,MAE,549,
+accuracy,Aus. Elec. Demand,VisionTS++,MAE,180.99,
+accuracy,Aus. Elec. Demand,VisionTS++,MAE,226.31,
+accuracy,Aus. Emission,VisionTS++,MAE,1,
+accuracy,Aus. Emission,VisionTS++,MAE,1,
+accuracy,Pedestrian Counts,VisionTS++,MAE,61.47,
+accuracy,Pedestrian Counts,VisionTS++,MAE,62.55,
+accuracy,Vehicle Trips,VisionTS++,MAE,20.67,
+accuracy,Vehicle Trips,VisionTS++,MAE,19.98,
+accuracy,KDD cup,VisionTS++,MAE,38.75,
+accuracy,KDD cup,VisionTS++,MAE,38.89,
+accuracy,Weather Low,VisionTS++,MAE,1.73,
+accuracy,Weather Low,VisionTS++,MAE,1.73,
+accuracy,NNS Daily,VisionTS++,MAE,3.51,
+accuracy,NNS Daily,VisionTS++,MAE,3.41,
+accuracy,NNS Weekly,VisionTS++,MAE,14.84,
+accuracy,NNS Weekly,VisionTS++,MAE,14.12,
+accuracy,Carparts,VisionTS++,MAE,0.44,
+accuracy,Carparts,VisionTS++,MAE,0.43,
+accuracy,FRED MID,VisionTS++,MAE,2722.75,
+accuracy,FRED MID,VisionTS++,MAE,2,
+accuracy,Traffic Hourly,VisionTS++,MAE,0.013,
+accuracy,Traffic Hourly,VisionTS++,MAE,0.016,
+accuracy,Traffic Weekly,VisionTS++,MAE,1.08,
+accuracy,Traffic Weekly,VisionTS++,MAE,1.07,
+accuracy,Rideshare,VisionTS++,MAE,1.37,
+accuracy,Rideshare,VisionTS++,MAE,1.36,
+accuracy,Hospital,VisionTS++,MAE,17.30,
+accuracy,Hospital,VisionTS++,MAE,17.00,
+accuracy,COVID Deaths,VisionTS++,MAE,114.97,
+accuracy,COVID Deaths,VisionTS++,MAE,151.53,
+accuracy,Temperature Rain,VisionTS++,MAE,4.83,
+accuracy,Temperature Rain,VisionTS++,MAE,5.17,
+accuracy,Sampost,VisionTS++,MAE,0.25,
+accuracy,Sampost,VisionTS++,MAE,0.19,
+accuracy,Saugeon River Phase,VisionTS++,MAE,23.24,
+accuracy,Saugeon River Phase,VisionTS++,MAE,24.24,
+accuracy,US Births,VisionTS++,MAE,420.22,
+accuracy,US Births,VisionTS++,MAE,411.48,
+accuracy,Normalized MAE,VisionTS++,MAE,0.544,
+accuracy,Normalized MAE,VisionTS++,MAE,0.553,
+accuracy,Electricity,VisionTS++_l,CRPS,0.041,
+accuracy,Electricity,VisionTS++_b,CRPS,0.042,
+accuracy,MASE,VisionTS++_l,CRPS,0.631,
+accuracy,MASE,VisionTS++_b,CRPS,0.755,
+accuracy,Solar,VisionTS++_l,CRPS,0.353,
+accuracy,Solar,VisionTS++_b,CRPS,0.353,
+accuracy,MASE,VisionTS++_l,CRPS,1.155,
+accuracy,MASE,VisionTS++_b,CRPS,1.141,
+accuracy,Walmart,VisionTS++_l,CRPS,0.061,
+accuracy,Walmart,VisionTS++_b,CRPS,0.064,
+accuracy,MASE,VisionTS++_l,CRPS,0.689,
+accuracy,MASE,VisionTS++_b,CRPS,0.949,
+accuracy,Weather,VisionTS++_l,CRPS,0.038,
+accuracy,Weather,VisionTS++_b,CRPS,0.038,
+accuracy,MASE,VisionTS++_l,CRPS,0.447,
+accuracy,MASE,VisionTS++_b,CRPS,0.737,
+accuracy,Istanbul Traffic,VisionTS++_l,CRPS,0.105,
+accuracy,Istanbul Traffic,VisionTS++_b,CRPS,0.115,
+accuracy,MASE,VisionTS++_l,CRPS,0.616,
+accuracy,MASE,VisionTS++_b,CRPS,0.706,
+accuracy,Turkey Power,VisionTS++_l,CRPS,0.038,
+accuracy,Turkey Power,VisionTS++_b,CRPS,0.036,
+accuracy,MASE,VisionTS++_l,CRPS,0.737,
+accuracy,MASE,VisionTS++_b,CRPS,0.856,
+accuracy,Norm.,VisionTS++_l,CRPS,0.506,
+accuracy,Norm.,VisionTS++_b,CRPS,0.515,
+accuracy,MASE,VisionTS++_l,CRPS,0.677,
+accuracy,MASE,VisionTS++_b,CRPS,0.838,
+accuracy,ETTm1,VisionTS++_l,MSE,0.354,
+accuracy,ETTm1,VisionTS++_b,MSE,0.360,
+accuracy,MAE,VisionTS++_l,MSE,0.372,
+accuracy,MAE,VisionTS++_b,MSE,0.372,
+accuracy,ETTm2,VisionTS++_l,MSE,0.244,
+accuracy,ETTm2,VisionTS++_b,MSE,0.244,
+accuracy,MAE,VisionTS++_l,MSE,0.298,
+accuracy,MAE,VisionTS++_b,MSE,0.321,
+accuracy,ETTh1,VisionTS++_l,MSE,0.403,
+accuracy,ETTh1,VisionTS++_b,MSE,0.402,
+accuracy,MAE,VisionTS++_l,MSE,0.416,
+accuracy,MAE,VisionTS++_b,MSE,0.414,
+accuracy,ETTh2,VisionTS++_l,MSE,0.327,
+accuracy,ETTh2,VisionTS++_b,MSE,0.333,
+accuracy,MAE,VisionTS++_l,MSE,0.370,
+accuracy,MAE,VisionTS++_b,MSE,0.375,
+accuracy,Electricity,VisionTS++_l,MSE,0.181,
+accuracy,Electricity,VisionTS++_b,MSE,0.184,
+accuracy,MAE,VisionTS++_l,MSE,0.265,
+accuracy,MAE,VisionTS++_b,MSE,0.294,
+accuracy,Weather,VisionTS++_l,MSE,0.226,
+accuracy,Weather,VisionTS++_b,MSE,0.222,
+accuracy,MAE,VisionTS++_l,MSE,0.241,
+accuracy,MAE,VisionTS++_b,MSE,0.292,
+accuracy,Average,VisionTS++_l,MSE,0.289,
+accuracy,Average,VisionTS++_b,MSE,0.291,
+accuracy,MAE,VisionTS++_l,MSE,0.327,
+accuracy,MAE,VisionTS++_b,MSE,0.345,
+accuracy,Monash,VisionTS++-b,MAE,0.553,
+accuracy,PF,VisionTS++-b,MAE,0.677,
+accuracy,CRPS,VisionTS++-b,MAE,0.627,
+accuracy,ETTm1,VisionTS++-b,MAE,0.360,
+accuracy,MAE,VisionTS++-b,MAE,0.396,
+accuracy,ETTm2,VisionTS++-b,MAE,0.244,
+accuracy,MAE,VisionTS++-b,MAE,0.337,
+accuracy,ETTh1,VisionTS++-b,MAE,0.402,
+accuracy,MAE,VisionTS++-b,MAE,0.45,
+accuracy,ETTh2,VisionTS++-b,MAE,0.333,
+accuracy,MAE,VisionTS++-b,MAE,0.439,
+accuracy,Electricity,VisionTS++-b,MAE,0.184,
+accuracy,MAE,VisionTS++-b,MAE,0.298,
+accuracy,Weather,VisionTS++-b,MAE,0.222,
+accuracy,MAE,VisionTS++-b,MAE,0.257,
diff --git a/result/per_paper/2508.04379/accuracy_efficiency_traced.csv b/result/per_paper/2508.04379/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eaa8b042e48ae51dfcc34864ca9b307c4405ffff
--- /dev/null
+++ b/result/per_paper/2508.04379/accuracy_efficiency_traced.csv
@@ -0,0 +1,132 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,M1 Monthly,VisionTS++,MAE,1919.97,,1,1,1
+accuracy,M1 Monthly,VisionTS++,MAE,1,,1,1,2
+accuracy,M3 Monthly,VisionTS++,MAE,591.44,,1,2,1
+accuracy,M3 Monthly,VisionTS++,MAE,581.68,,1,2,2
+accuracy,M3 Other,VisionTS++,MAE,180.99,,1,3,1
+accuracy,M3 Other,VisionTS++,MAE,186.13,,1,3,2
+accuracy,M4 Monthly,VisionTS++,MAE,533.16,,1,4,1
+accuracy,M4 Monthly,VisionTS++,MAE,533.74,,1,4,2
+accuracy,M4 Weekly,VisionTS++,MAE,281.76,,1,5,1
+accuracy,M4 Weekly,VisionTS++,MAE,280.88,,1,5,2
+accuracy,M4 Daily,VisionTS++,MAE,190.54,,1,6,1
+accuracy,M4 Daily,VisionTS++,MAE,172.31,,1,6,2
+accuracy,M4 Hourly,VisionTS++,MAE,169.17,,1,7,1
+accuracy,M4 Hourly,VisionTS++,MAE,202.99,,1,7,2
+accuracy,Tourism Quarterly,VisionTS++,MAE,5623.23,,1,8,1
+accuracy,Tourism Quarterly,VisionTS++,MAE,6,,1,8,2
+accuracy,Tourism Monthly,VisionTS++,MAE,1667.65,,1,9,1
+accuracy,Tourism Monthly,VisionTS++,MAE,2,,1,9,2
+accuracy,CIF 2016,VisionTS++,MAE,5664488.37,,1,10,1
+accuracy,CIF 2016,VisionTS++,MAE,549,,1,10,2
+accuracy,Aus. Elec. Demand,VisionTS++,MAE,180.99,,1,11,1
+accuracy,Aus. Elec. Demand,VisionTS++,MAE,226.31,,1,11,2
+accuracy,Aus. Emission,VisionTS++,MAE,1,,1,12,1
+accuracy,Aus. Emission,VisionTS++,MAE,1,,1,12,2
+accuracy,Pedestrian Counts,VisionTS++,MAE,61.47,,1,13,1
+accuracy,Pedestrian Counts,VisionTS++,MAE,62.55,,1,13,2
+accuracy,Vehicle Trips,VisionTS++,MAE,20.67,,1,14,1
+accuracy,Vehicle Trips,VisionTS++,MAE,19.98,,1,14,2
+accuracy,KDD cup,VisionTS++,MAE,38.75,,1,15,1
+accuracy,KDD cup,VisionTS++,MAE,38.89,,1,15,2
+accuracy,Weather Low,VisionTS++,MAE,1.73,,1,16,1
+accuracy,Weather Low,VisionTS++,MAE,1.73,,1,16,2
+accuracy,NNS Daily,VisionTS++,MAE,3.51,,1,17,1
+accuracy,NNS Daily,VisionTS++,MAE,3.41,,1,17,2
+accuracy,NNS Weekly,VisionTS++,MAE,14.84,,1,18,1
+accuracy,NNS Weekly,VisionTS++,MAE,14.12,,1,18,2
+accuracy,Carparts,VisionTS++,MAE,0.44,,1,19,1
+accuracy,Carparts,VisionTS++,MAE,0.43,,1,19,2
+accuracy,FRED MID,VisionTS++,MAE,2722.75,,1,20,1
+accuracy,FRED MID,VisionTS++,MAE,2,,1,20,2
+accuracy,Traffic Hourly,VisionTS++,MAE,0.013,,1,21,1
+accuracy,Traffic Hourly,VisionTS++,MAE,0.016,,1,21,2
+accuracy,Traffic Weekly,VisionTS++,MAE,1.08,,1,22,1
+accuracy,Traffic Weekly,VisionTS++,MAE,1.07,,1,22,2
+accuracy,Rideshare,VisionTS++,MAE,1.37,,1,23,1
+accuracy,Rideshare,VisionTS++,MAE,1.36,,1,23,2
+accuracy,Hospital,VisionTS++,MAE,17.30,,1,24,1
+accuracy,Hospital,VisionTS++,MAE,17.00,,1,24,2
+accuracy,COVID Deaths,VisionTS++,MAE,114.97,,1,25,1
+accuracy,COVID Deaths,VisionTS++,MAE,151.53,,1,25,2
+accuracy,Temperature Rain,VisionTS++,MAE,4.83,,1,26,1
+accuracy,Temperature Rain,VisionTS++,MAE,5.17,,1,26,2
+accuracy,Sampost,VisionTS++,MAE,0.25,,1,27,1
+accuracy,Sampost,VisionTS++,MAE,0.19,,1,27,2
+accuracy,Saugeon River Phase,VisionTS++,MAE,23.24,,1,28,1
+accuracy,Saugeon River Phase,VisionTS++,MAE,24.24,,1,28,2
+accuracy,US Births,VisionTS++,MAE,420.22,,1,29,1
+accuracy,US Births,VisionTS++,MAE,411.48,,1,29,2
+accuracy,Normalized MAE,VisionTS++,MAE,0.544,,1,30,1
+accuracy,Normalized MAE,VisionTS++,MAE,0.553,,1,30,2
+accuracy,Electricity,VisionTS++_l,CRPS,0.041,,2,2,2
+accuracy,Electricity,VisionTS++_b,CRPS,0.042,,2,2,3
+accuracy,MASE,VisionTS++_l,CRPS,0.631,,2,3,2
+accuracy,MASE,VisionTS++_b,CRPS,0.755,,2,3,3
+accuracy,Solar,VisionTS++_l,CRPS,0.353,,2,4,2
+accuracy,Solar,VisionTS++_b,CRPS,0.353,,2,4,3
+accuracy,MASE,VisionTS++_l,CRPS,1.155,,2,5,2
+accuracy,MASE,VisionTS++_b,CRPS,1.141,,2,5,3
+accuracy,Walmart,VisionTS++_l,CRPS,0.061,,2,6,2
+accuracy,Walmart,VisionTS++_b,CRPS,0.064,,2,6,3
+accuracy,MASE,VisionTS++_l,CRPS,0.689,,2,7,2
+accuracy,MASE,VisionTS++_b,CRPS,0.949,,2,7,3
+accuracy,Weather,VisionTS++_l,CRPS,0.038,,2,8,2
+accuracy,Weather,VisionTS++_b,CRPS,0.038,,2,8,3
+accuracy,MASE,VisionTS++_l,CRPS,0.447,,2,9,2
+accuracy,MASE,VisionTS++_b,CRPS,0.737,,2,9,3
+accuracy,Istanbul Traffic,VisionTS++_l,CRPS,0.105,,2,10,2
+accuracy,Istanbul Traffic,VisionTS++_b,CRPS,0.115,,2,10,3
+accuracy,MASE,VisionTS++_l,CRPS,0.616,,2,11,2
+accuracy,MASE,VisionTS++_b,CRPS,0.706,,2,11,3
+accuracy,Turkey Power,VisionTS++_l,CRPS,0.038,,2,12,2
+accuracy,Turkey Power,VisionTS++_b,CRPS,0.036,,2,12,3
+accuracy,MASE,VisionTS++_l,CRPS,0.737,,2,13,2
+accuracy,MASE,VisionTS++_b,CRPS,0.856,,2,13,3
+accuracy,Norm.,VisionTS++_l,CRPS,0.506,,2,14,2
+accuracy,Norm.,VisionTS++_b,CRPS,0.515,,2,14,3
+accuracy,MASE,VisionTS++_l,CRPS,0.677,,2,15,2
+accuracy,MASE,VisionTS++_b,CRPS,0.838,,2,15,3
+accuracy,ETTm1,VisionTS++_l,MSE,0.354,,3,2,2
+accuracy,ETTm1,VisionTS++_b,MSE,0.360,,3,2,3
+accuracy,MAE,VisionTS++_l,MSE,0.372,,3,3,2
+accuracy,MAE,VisionTS++_b,MSE,0.372,,3,3,3
+accuracy,ETTm2,VisionTS++_l,MSE,0.244,,3,4,2
+accuracy,ETTm2,VisionTS++_b,MSE,0.244,,3,4,3
+accuracy,MAE,VisionTS++_l,MSE,0.298,,3,5,2
+accuracy,MAE,VisionTS++_b,MSE,0.321,,3,5,3
+accuracy,ETTh1,VisionTS++_l,MSE,0.403,,3,6,2
+accuracy,ETTh1,VisionTS++_b,MSE,0.402,,3,6,3
+accuracy,MAE,VisionTS++_l,MSE,0.416,,3,7,2
+accuracy,MAE,VisionTS++_b,MSE,0.414,,3,7,3
+accuracy,ETTh2,VisionTS++_l,MSE,0.327,,3,8,2
+accuracy,ETTh2,VisionTS++_b,MSE,0.333,,3,8,3
+accuracy,MAE,VisionTS++_l,MSE,0.370,,3,9,2
+accuracy,MAE,VisionTS++_b,MSE,0.375,,3,9,3
+accuracy,Electricity,VisionTS++_l,MSE,0.181,,3,10,2
+accuracy,Electricity,VisionTS++_b,MSE,0.184,,3,10,3
+accuracy,MAE,VisionTS++_l,MSE,0.265,,3,11,2
+accuracy,MAE,VisionTS++_b,MSE,0.294,,3,11,3
+accuracy,Weather,VisionTS++_l,MSE,0.226,,3,12,2
+accuracy,Weather,VisionTS++_b,MSE,0.222,,3,12,3
+accuracy,MAE,VisionTS++_l,MSE,0.241,,3,13,2
+accuracy,MAE,VisionTS++_b,MSE,0.292,,3,13,3
+accuracy,Average,VisionTS++_l,MSE,0.289,,3,14,2
+accuracy,Average,VisionTS++_b,MSE,0.291,,3,14,3
+accuracy,MAE,VisionTS++_l,MSE,0.327,,3,15,2
+accuracy,MAE,VisionTS++_b,MSE,0.345,,3,15,3
+accuracy,Monash,VisionTS++-b,MAE,0.553,,5,1,2
+accuracy,PF,VisionTS++-b,MAE,0.677,,5,2,2
+accuracy,CRPS,VisionTS++-b,MAE,0.627,,5,3,2
+accuracy,ETTm1,VisionTS++-b,MAE,0.360,,5,4,2
+accuracy,MAE,VisionTS++-b,MAE,0.396,,5,5,2
+accuracy,ETTm2,VisionTS++-b,MAE,0.244,,5,6,2
+accuracy,MAE,VisionTS++-b,MAE,0.337,,5,7,2
+accuracy,ETTh1,VisionTS++-b,MAE,0.402,,5,8,2
+accuracy,MAE,VisionTS++-b,MAE,0.45,,5,9,2
+accuracy,ETTh2,VisionTS++-b,MAE,0.333,,5,10,2
+accuracy,MAE,VisionTS++-b,MAE,0.439,,5,11,2
+accuracy,Electricity,VisionTS++-b,MAE,0.184,,5,12,2
+accuracy,MAE,VisionTS++-b,MAE,0.298,,5,13,2
+accuracy,Weather,VisionTS++-b,MAE,0.222,,5,14,2
+accuracy,MAE,VisionTS++-b,MAE,0.257,,5,15,2
diff --git a/result/per_paper/2508.04379/components_architecture.csv b/result/per_paper/2508.04379/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f65de25890ac940ab10409254b73dd37358fd7b7
--- /dev/null
+++ b/result/per_paper/2508.04379/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Vision-Model-Based Filtering,"A mechanism to identify and discard low-quality time series samples by leveraging the vision model's constraints, enhancing pre-training stability and mitigating modality gaps.",proposed_here,"Shen et al., 2025",we introduce a filtering mechanism that leverages the vision model itself to select high-quality time series. we identify and discard samples with out-of-range values or abrupt anomalies—inputs incompatible with the model’s constraints.
+Colorized Multivariate Conversion,"A method to encode multivariate time series as multi-subfigure RGB images, mapping each variate to a distinct subfigure to capture cross-variate dependencies as spatial relationships.",proposed_here,"Shen et al., 2025","we encode multivariate time series as multi-subfigure RGB images, where each variate is mapped to a distinct subfigure. This allows cross-variate dependencies to be better captured as spatial relationships between subfigures."
+Multi-Quantile Forecasting,"A technique using parallel reconstruction heads to generate quantile forecasts without parametric assumptions, enabling probabilistic predictions for time series forecasting.",proposed_here,"Shen et al., 2025","multi-quantile forecasting, using parallel reconstruction heads to generate quantile forecasts without parametric assumptions."
+Continual Pre-Training (CPT) on Time Series,A training paradigm where the vision model is continually pre-trained on large-scale time series data to adapt its architecture for cross-modal transfer.,proposed_here,"Shen et al., 2025","VISIONTS++, a TSFM based on continual pre-training of a vision model on large-scale time series."
diff --git a/result/per_paper/2508.04379/computational.csv b/result/per_paper/2508.04379/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..073392b8e1c90e807352a44b985e74a718a2ba29
--- /dev/null
+++ b/result/per_paper/2508.04379/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No explicit mention of hardware type (e.g., GPU/TPU) in the text."
+num_devices,,,not_reported,"No explicit mention of the number of devices (e.g., GPUs) used during training or inference."
+training_cost,,,not_reported,No explicit mention of monetary or resource cost for training.
+training_batch_size,512,,stated,"Training runs for 100,000 steps with a batch size of 512."
+training_steps_or_epochs,100000,steps,stated,"Continual pre-training runs for 100,000 steps."
+precision,,,not_reported,"No explicit mention of training precision (e.g., 32-bit/16-bit)."
+inference_latency,,,not_reported,No explicit mention of inference latency.
+inference_throughput,,,not_reported,No explicit mention of inference throughput.
+peak_memory,,,not_reported,No explicit mention of peak memory usage.
+flops_or_macs,,,not_reported,No explicit mention of FLOPs or MACs.
+num_inference_samples,,,not_reported,No explicit mention of the number of inference samples evaluated.
+params,112000000,parameters,stated,Based on MAE (base) architecture with 112M parameters for VisionTS++_base and 330M for VisionTS++_large.
+context_lengths_evaluated,,,not_reported,No explicit mention of specific context lengths (L) evaluated.
+horizon_lengths_evaluated,,,not_reported,"Horizon lengths (T) are dataset-specific (e.g., 96, 192, 336, 720 in tables), but not explicitly stated as model-specific parameters."
+inference_batch_size,,,not_reported,No explicit mention of inference batch size.
diff --git a/result/per_paper/2509.21190/accuracy_efficiency.csv b/result/per_paper/2509.21190/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a697ce3b2c8c4052fb1defac23ad9711fe6cb295
--- /dev/null
+++ b/result/per_paper/2509.21190/accuracy_efficiency.csv
@@ -0,0 +1,65 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,$DADA^†$,TimeRCD,accuracy,89.37,
+accuracy,$MOMENT^†$,TimeRCD,accuracy,87.54,
+accuracy,TimesFM,TimeRCD,accuracy,81.88,
+accuracy,Chronos,TimeRCD,accuracy,90.12,
+accuracy,Time MOE,TimeRCD,accuracy,76.34,
+accuracy,MovingVar.,TimeRCD,accuracy,83.90,
+accuracy,$DADA^†$,TimeRCD,accuracy,42.50,
+accuracy,$MOMENT^†$,TimeRCD,accuracy,33.15,
+accuracy,TimesFM,TimeRCD,accuracy,48.95,
+accuracy,Chronos,TimeRCD,accuracy,45.45,
+accuracy,Time MOE,TimeRCD,accuracy,25.95,
+accuracy,MovingVar.,TimeRCD,accuracy,22.53,
+accuracy,$DADA^†$,TimeRCD,accuracy,32.76,
+accuracy,$MOMENT^†$,TimeRCD,accuracy,30.69,
+accuracy,TimesFM,TimeRCD,accuracy,34.28,
+accuracy,Chronos,TimeRCD,accuracy,32.69,
+accuracy,Time MOE,TimeRCD,accuracy,26.52,
+accuracy,MovingVar.,TimeRCD,accuracy,22.74,
+accuracy,$DADA^†$,TimeRCD,accuracy,24.97,
+accuracy,$MOMENT^†$,TimeRCD,accuracy,37.35,
+accuracy,TimesFM,TimeRCD,accuracy,19.56,
+accuracy,Chronos,TimeRCD,accuracy,19.00,
+accuracy,Time MOE,TimeRCD,accuracy,16.63,
+accuracy,MovingVar.,TimeRCD,accuracy,22.47,
+accuracy,TranAD,TimeRCD,accuracy,83.19,
+accuracy,USAD,TimeRCD,accuracy,71.08,
+accuracy,OmniAnomaly,TimeRCD,accuracy,80.32,
+accuracy,LOF,TimeRCD,accuracy,81.06,
+accuracy,IForest,TimeRCD,accuracy,52.81,
+accuracy,Sub-PCA,TimeRCD,accuracy,75.39,
+accuracy,DCdetector,TimeRCD,accuracy,71.83,
+accuracy,TFMAE,TimeRCD,accuracy,78.25,
+accuracy,TranAD,TimeRCD,accuracy,22.63,
+accuracy,USAD,TimeRCD,accuracy,20.99,
+accuracy,OmniAnomaly,TimeRCD,accuracy,51.17,
+accuracy,LOF,TimeRCD,accuracy,27.97,
+accuracy,IForest,TimeRCD,accuracy,7.64,
+accuracy,Sub-PCA,TimeRCD,accuracy,32.75,
+accuracy,DCdetector,TimeRCD,accuracy,6.61,
+accuracy,TFMAE,TimeRCD,accuracy,19.41,
+accuracy,TranAD,TimeRCD,accuracy,34.85,
+accuracy,USAD,TimeRCD,accuracy,30.66,
+accuracy,OmniAnomaly,TimeRCD,accuracy,47.05,
+accuracy,LOF,TimeRCD,accuracy,30.28,
+accuracy,IForest,TimeRCD,accuracy,8.37,
+accuracy,Sub-PCA,TimeRCD,accuracy,33.96,
+accuracy,DCdetector,TimeRCD,accuracy,5.19,
+accuracy,TFMAE,TimeRCD,accuracy,9.48,
+accuracy,F1_T,Affiliation-F,Accuracy,1,
+accuracy,F1_T,2k,Accuracy,1,
+accuracy,Standard-F1,Affiliation-F,Accuracy,1,
+accuracy,Standard-F1,2k,Accuracy,1,
+accuracy,F1_T,Affiliation-F,accuracy,1,
+accuracy,F1_T,700M,accuracy,1,
+accuracy,Standard-F1,Affiliation-F,accuracy,1,
+accuracy,Standard-F1,700M,accuracy,1,
+accuracy,Point,TimeRCD,F1,0.912,
+accuracy,$F1_T$,TimeRCD,F1,0.441,
+accuracy,Standard-F1,TimeRCD,F1,0.352,
+accuracy,VUS-PR,TimeRCD,F1,0.293,
+accuracy,Contextual,TimeRCD,F1,0.949,
+accuracy,$F1_T$,TimeRCD,F1,0.174,
+accuracy,Standard-F1,TimeRCD,F1,0.154,
+accuracy,VUS-PR,TimeRCD,F1,0.108,
diff --git a/result/per_paper/2509.21190/accuracy_efficiency_traced.csv b/result/per_paper/2509.21190/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4bcd27c6bde1ba828a2f6583b3d1616357902603
--- /dev/null
+++ b/result/per_paper/2509.21190/accuracy_efficiency_traced.csv
@@ -0,0 +1,65 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,$DADA^†$,TimeRCD,accuracy,89.37,,0,4,1
+accuracy,$MOMENT^†$,TimeRCD,accuracy,87.54,,0,5,1
+accuracy,TimesFM,TimeRCD,accuracy,81.88,,0,6,1
+accuracy,Chronos,TimeRCD,accuracy,90.12,,0,7,1
+accuracy,Time MOE,TimeRCD,accuracy,76.34,,0,8,1
+accuracy,MovingVar.,TimeRCD,accuracy,83.90,,0,9,1
+accuracy,$DADA^†$,TimeRCD,accuracy,42.50,,0,11,1
+accuracy,$MOMENT^†$,TimeRCD,accuracy,33.15,,0,12,1
+accuracy,TimesFM,TimeRCD,accuracy,48.95,,0,13,1
+accuracy,Chronos,TimeRCD,accuracy,45.45,,0,14,1
+accuracy,Time MOE,TimeRCD,accuracy,25.95,,0,15,1
+accuracy,MovingVar.,TimeRCD,accuracy,22.53,,0,16,1
+accuracy,$DADA^†$,TimeRCD,accuracy,32.76,,0,18,1
+accuracy,$MOMENT^†$,TimeRCD,accuracy,30.69,,0,19,1
+accuracy,TimesFM,TimeRCD,accuracy,34.28,,0,20,1
+accuracy,Chronos,TimeRCD,accuracy,32.69,,0,21,1
+accuracy,Time MOE,TimeRCD,accuracy,26.52,,0,22,1
+accuracy,MovingVar.,TimeRCD,accuracy,22.74,,0,23,1
+accuracy,$DADA^†$,TimeRCD,accuracy,24.97,,0,25,1
+accuracy,$MOMENT^†$,TimeRCD,accuracy,37.35,,0,26,1
+accuracy,TimesFM,TimeRCD,accuracy,19.56,,0,27,1
+accuracy,Chronos,TimeRCD,accuracy,19.00,,0,28,1
+accuracy,Time MOE,TimeRCD,accuracy,16.63,,0,29,1
+accuracy,MovingVar.,TimeRCD,accuracy,22.47,,0,30,1
+accuracy,TranAD,TimeRCD,accuracy,83.19,,0,34,1
+accuracy,USAD,TimeRCD,accuracy,71.08,,0,35,1
+accuracy,OmniAnomaly,TimeRCD,accuracy,80.32,,0,36,1
+accuracy,LOF,TimeRCD,accuracy,81.06,,0,37,1
+accuracy,IForest,TimeRCD,accuracy,52.81,,0,38,1
+accuracy,Sub-PCA,TimeRCD,accuracy,75.39,,0,39,1
+accuracy,DCdetector,TimeRCD,accuracy,71.83,,0,40,1
+accuracy,TFMAE,TimeRCD,accuracy,78.25,,0,41,1
+accuracy,TranAD,TimeRCD,accuracy,22.63,,0,43,1
+accuracy,USAD,TimeRCD,accuracy,20.99,,0,44,1
+accuracy,OmniAnomaly,TimeRCD,accuracy,51.17,,0,45,1
+accuracy,LOF,TimeRCD,accuracy,27.97,,0,46,1
+accuracy,IForest,TimeRCD,accuracy,7.64,,0,47,1
+accuracy,Sub-PCA,TimeRCD,accuracy,32.75,,0,48,1
+accuracy,DCdetector,TimeRCD,accuracy,6.61,,0,49,1
+accuracy,TFMAE,TimeRCD,accuracy,19.41,,0,50,1
+accuracy,TranAD,TimeRCD,accuracy,34.85,,0,52,1
+accuracy,USAD,TimeRCD,accuracy,30.66,,0,53,1
+accuracy,OmniAnomaly,TimeRCD,accuracy,47.05,,0,54,1
+accuracy,LOF,TimeRCD,accuracy,30.28,,0,55,1
+accuracy,IForest,TimeRCD,accuracy,8.37,,0,56,1
+accuracy,Sub-PCA,TimeRCD,accuracy,33.96,,0,57,1
+accuracy,DCdetector,TimeRCD,accuracy,5.19,,0,58,1
+accuracy,TFMAE,TimeRCD,accuracy,9.48,,0,59,1
+accuracy,F1_T,Affiliation-F,Accuracy,1,,1,17,0
+accuracy,F1_T,2k,Accuracy,1,,1,17,0
+accuracy,Standard-F1,Affiliation-F,Accuracy,1,,1,32,0
+accuracy,Standard-F1,2k,Accuracy,1,,1,32,0
+accuracy,F1_T,Affiliation-F,accuracy,1,,2,5,0
+accuracy,F1_T,700M,accuracy,1,,2,5,0
+accuracy,Standard-F1,Affiliation-F,accuracy,1,,2,8,0
+accuracy,Standard-F1,700M,accuracy,1,,2,8,0
+accuracy,Point,TimeRCD,F1,0.912,,3,1,2
+accuracy,$F1_T$,TimeRCD,F1,0.441,,3,2,2
+accuracy,Standard-F1,TimeRCD,F1,0.352,,3,3,2
+accuracy,VUS-PR,TimeRCD,F1,0.293,,3,4,2
+accuracy,Contextual,TimeRCD,F1,0.949,,3,5,2
+accuracy,$F1_T$,TimeRCD,F1,0.174,,3,6,2
+accuracy,Standard-F1,TimeRCD,F1,0.154,,3,7,2
+accuracy,VUS-PR,TimeRCD,F1,0.108,,3,8,2
diff --git a/result/per_paper/2509.21190/components_architecture.csv b/result/per_paper/2509.21190/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e6115a3b6f87ff87b85855864aa5bfb84de87521
--- /dev/null
+++ b/result/per_paper/2509.21190/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Relative Context Discrepancy (RCD),"A pre-training paradigm that trains the model to detect anomalies by comparing a query pattern with its surrounding context, enabling inference of normality from input context rather than fixed global patterns.",proposed_here,,"We introduce TimeRCD, a foundation model for TSAD built on Relative Context Discrepancy (RCD), a pre-training paradigm that trains the model to detect anomalies by comparing a query pattern with its surrounding context."
+Transformer Architecture,"A standard Transformer architecture used to implement the relational formulation of RCD, enabling context-aware anomaly detection.",reused_cited,"Vaswani et al., 2017","This relational formulation, implemented with a standard Transformer architecture, enables the model to infer normality from the input context."
+Synthetic Corpus,"A large-scale synthetic corpus with context-dependent anomaly labels, constructed to provide supervised pre-training signals for RCD.",proposed_here,,We further construct a large-scale synthetic corpus with context-dependent anomaly labels to provide supervised pre-training signals for RCD.
+Anomaly Scoring Mechanism,A method to compute anomaly scores by comparing a query pattern with its surrounding context using attention mechanisms within the Transformer.,proposed_here,,RCD by attention -.-> Anomalous RCD (Anomalous RCD) [from diagram].
diff --git a/result/per_paper/2509.21190/computational.csv b/result/per_paper/2509.21190/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2509.21190/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2509.25826/accuracy_efficiency.csv b/result/per_paper/2509.25826/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5686d22be7fcf4657174b04dae50918d353868be
--- /dev/null
+++ b/result/per_paper/2509.25826/accuracy_efficiency.csv
@@ -0,0 +1,8 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,#Params,Kairos_s,MASE,23,
+accuracy,#Params,Kairos_b,MASE,53,
+accuracy,MASE ↓,Kairos_s,MASE,0.748,
+accuracy,MASE ↓,Kairos_b,MASE,0.738,
+accuracy,CRPS ↓,Kairos_s,MASE,0.554,
+accuracy,CRPS ↓,Kairos_b,MASE,0.548,
+accuracy,MASE ↓,DRoPE (Ours),MASE,0.738,
diff --git a/result/per_paper/2509.25826/accuracy_efficiency_traced.csv b/result/per_paper/2509.25826/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eea2e49e06da6f4f1fcfb4ff71d1629a79afa2ec
--- /dev/null
+++ b/result/per_paper/2509.25826/accuracy_efficiency_traced.csv
@@ -0,0 +1,8 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,#Params,Kairos_s,MASE,23,,1,2,13
+accuracy,#Params,Kairos_b,MASE,53,,1,2,14
+accuracy,MASE ↓,Kairos_s,MASE,0.748,,1,3,13
+accuracy,MASE ↓,Kairos_b,MASE,0.738,,1,3,14
+accuracy,CRPS ↓,Kairos_s,MASE,0.554,,1,4,13
+accuracy,CRPS ↓,Kairos_b,MASE,0.548,,1,4,14
+accuracy,MASE ↓,DRoPE (Ours),MASE,0.738,,7,1,1
diff --git a/result/per_paper/2509.25826/components_architecture.csv b/result/per_paper/2509.25826/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..51de418053575f90bffedf1876913d89debf368a
--- /dev/null
+++ b/result/per_paper/2509.25826/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Mixture-of-Size Encoder,"Adaptive encoder that models time series at multiple granularities based on local characteristics, inspired by null experts in Mixture-of-Experts (MoE).",proposed_here,"Feng et al., 2023","Inspired by null experts in Mixture-of-Experts [43, 20], we introduce computation-free null experts to dynamically adjust the number of active granularities."
+Dynamic Patching Tokenizer,"Tokenization method that dynamically adapts to local information density, enabling fine-grained temporal abstraction without increasing model width or depth.",proposed_here,,"Our KAIROS uses Mixture-of-Size Tokenization to dynamically adapt to information density, enabling the learning of generalizable rules rather than memorization."
+Multi-Granularity Positional Embedding,"Positional encoding mechanism based on dynamic rotary encodings, conditioning on instance-level spectral features and temporal structure induced by dynamic patching tokenization.",proposed_here,"Feng et al., 2023","We design a multi-granularity positional embedding based on dynamic rotary encodings, which conditions on instance-level spectral features and temporal structure induced by dynamic patching tokenization."
+Dynamic Rotary Position Embedding (DRoPE),Enhanced positional encoding that modulates temporal scales using spectral features and granularity to generate instance-specific positional encodings.,proposed_here,"Feng et al., 2023",DRoPE modulates temporal scales using spectral features and granularity to generate instance-specific positional encodings.
+Multi-Patch Decoder,"Decoder that uses learnable tokens to predict future patches in parallel, mitigating cumulative errors in autoregressive generation.",proposed_here,"Feng et al., 2023","The Multi-Patch Decoder uses learnable tokens to predict future patches in parallel, which mitigates cumulative errors in autoregressive generation."
+Heterogeneity-Aware Transformer,Transformer architecture enhanced with DRoPE to model inter-patch dependencies while accounting for heterogeneous temporal dynamics.,proposed_here,"Feng et al., 2023",We enhance the Transformer with Dynamic Rotary Position Embedding (DRoPE) to model inter-patch dependencies.
diff --git a/result/per_paper/2509.25826/computational.csv b/result/per_paper/2509.25826/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5d6d0c991b45832235e04e9a0f251dda24f64227
--- /dev/null
+++ b/result/per_paper/2509.25826/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA TITAN RTX GPU,,stated,experimental setup involving an NVIDIA TITAN RTX GPU
+num_devices,1,,stated,single GPU in experimental setup
+training_cost,,,not_reported,not mentioned in text
+training_batch_size,,,not_reported,not mentioned in text
+training_steps_or_epochs,,,not_reported,not mentioned in text
+precision,,,not_reported,not mentioned in text
+inference_latency,0.061,s,stated,Table 6: KAIROS-Base inference time of 0.061s
+inference_throughput,,,not_reported,not mentioned in text
+peak_memory,,,not_reported,not mentioned in text
+flops_or_macs,,,not_reported,not mentioned in text
+num_inference_samples,,,not_reported,not mentioned in text
+params,53,M,stated,KAIROS-Base has 53M parameters in tables
+context_lengths_evaluated,2048,,stated,last table: Kairos context length of 2048
+horizon_lengths_evaluated,96,,stated,experimental setup: prediction horizon of 96
+inference_batch_size,1,,stated,single-batch inference speed benchmarked
diff --git a/result/per_paper/2509.26468/accuracy_efficiency.csv b/result/per_paper/2509.26468/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..1eda52992c1656357595ff8b3cbe7a9f12e7f29c
--- /dev/null
+++ b/result/per_paper/2509.26468/accuracy_efficiency.csv
@@ -0,0 +1,9 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Chronos-2,fev-bench,Avg. win rate (%),-2,
+accuracy,TimesFM-2.5,fev-bench,Avg. win rate (%),-2.5,
+accuracy,Toto-1.0,fev-bench,Avg. win rate (%),-1.0,
+accuracy,Moirai-2.0,fev-bench,Avg. win rate (%),-2.0,
+accuracy,Chronos-2,fev-bench,Avg. win rate (%),-2,
+accuracy,TimesFM-2.5,fev-bench,Avg. win rate (%),-2.5,
+accuracy,Toto-1.0,fev-bench,Avg. win rate (%),-1.0,
+accuracy,Moirai-2.0,fev-bench,Avg. win rate (%),-2.0,
diff --git a/result/per_paper/2509.26468/accuracy_efficiency_traced.csv b/result/per_paper/2509.26468/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..d0c391779e84c15bdbe6f128146f18b5fdd9ad62
--- /dev/null
+++ b/result/per_paper/2509.26468/accuracy_efficiency_traced.csv
@@ -0,0 +1,9 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Chronos-2,fev-bench,Avg. win rate (%),-2,,0,1,0
+accuracy,TimesFM-2.5,fev-bench,Avg. win rate (%),-2.5,,0,3,0
+accuracy,Toto-1.0,fev-bench,Avg. win rate (%),-1.0,,0,4,0
+accuracy,Moirai-2.0,fev-bench,Avg. win rate (%),-2.0,,0,5,0
+accuracy,Chronos-2,fev-bench,Avg. win rate (%),-2,,1,1,0
+accuracy,TimesFM-2.5,fev-bench,Avg. win rate (%),-2.5,,1,3,0
+accuracy,Toto-1.0,fev-bench,Avg. win rate (%),-1.0,,1,4,0
+accuracy,Moirai-2.0,fev-bench,Avg. win rate (%),-2.0,,1,5,0
diff --git a/result/per_paper/2509.26468/components_architecture.csv b/result/per_paper/2509.26468/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e764d62fcde00ad88c51e8f563facce5392b0c21
--- /dev/null
+++ b/result/per_paper/2509.26468/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+fev-bench,"A benchmark containing 100 forecasting tasks across seven real-world domains, including 46 tasks with covariates.",proposed_here,,"We introduce fev-bench, a forecast evaluation benchmark containing 100 tasks spanning 7 real-world application domains."
+fev,"A lightweight Python library for forecasting evaluation, emphasizing reproducibility and integration with existing workflows.",proposed_here,,"We introduce fev, a lightweight Python package for forecasting evaluation that introduces minimal dependencies while remaining compatible with popular forecasting libraries."
+MASE (Mean Absolute Scaled Error),"A scale-free metric for evaluating point forecasts, normalized by historical seasonal error.",reused_cited,"Hyndman & Koehler, 2006","The error for each series n and dimension d is normalized by the historical seasonal error $a_{n,d} = \frac{1}{T-m} \sum_{t=m+1}^{T} |y_{n,d,t} - y_{n,d,t-m}|$."
+SQL (Scaled Quantile Loss),"A probabilistic metric extending MASE's scale-independence, used for evaluating probabilistic forecasts.",reused_cited,"Makridakis et al., 2020; 2022","We adopt the Scaled Quantile Loss (SQL) as our primary probabilistic metric, since it is the natural extension of MASE."
+Bootstrapped Confidence Intervals,A principled aggregation method to quantify statistical significance of performance differences across models.,proposed_here,,fev-bench employs principled aggregation with bootstrapped confidence intervals to report performance along two dimensions: win rates and skill scores.
+Win Rates and Skill Scores,Two dimensions of performance reporting used to assess model effectiveness and robustness.,proposed_here,,fev-bench employs principled aggregation with bootstrapped confidence intervals to report performance along two dimensions: win rates and skill scores.
diff --git a/result/per_paper/2509.26468/computational.csv b/result/per_paper/2509.26468/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..10f6ed4936c8692ab0a6ea40ed9125e98d3bf753
--- /dev/null
+++ b/result/per_paper/2509.26468/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA A10G GPU,,stated,consumer hardware (single NVIDIA A10G GPU with 24GB RAM)
+num_devices,1,,stated,single NVIDIA A10G GPU
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2510.03911/accuracy_efficiency.csv b/result/per_paper/2510.03911/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2510.03911/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2510.03911/accuracy_efficiency_traced.csv b/result/per_paper/2510.03911/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2510.03911/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2510.03911/components_architecture.csv b/result/per_paper/2510.03911/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..00500b0ea395d5f915e8f38d53488665643ede9c
--- /dev/null
+++ b/result/per_paper/2510.03911/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+Chronos Foundation Model Encoder,"The encoder component of the Chronos time series foundation model, used to generate embeddings for time series data.",reused_cited,"Ansari et al., 2024",THEMIS extracts embeddings from the encoder of the Chronos time series foundation model.
+Self-Similarity Matrix,"A matrix representing pairwise similarities between time series segments, used as input for spectral decomposition-based outlier detection.",reused_cited,"Yue et al., 2022",Spectral Decomposition on the self-similarity matrix is applied to detect anomalies.
+Local Outlier Factor (LOF) Algorithm,An unsupervised outlier detection algorithm that identifies anomalies based on local density deviations in the embedding space.,reused_cited,"Breunig et al., 2000",Local Outlier Factor (LOF) is used to detect local outliers in the embeddings.
+Spectral Decomposition Method,A technique for analyzing the self-similarity matrix to identify anomalous patterns through frequency domain analysis.,reused_cited,"Yue et al., 2022",Spectral Decomposition on the self-similarity matrix is applied to detect anomalies.
+THEMIS Framework,A modular anomaly detection framework that combines pretrained Chronos embeddings with outlier detection algorithms for time series analysis.,proposed_here,,THEMIS is introduced as a novel framework that strategically uses pretrained knowledge from TSFMs for anomaly detection.
diff --git a/result/per_paper/2510.03911/computational.csv b/result/per_paper/2510.03911/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ce0515d0c63b96b223ec6e5a9ae676a2928bae5b
--- /dev/null
+++ b/result/per_paper/2510.03911/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,512,samples,stated,segment the input using fixed-length sliding windows of size L = 512
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2510.15821/accuracy_efficiency.csv b/result/per_paper/2510.15821/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..36b0d919f0e16593a927afa6894545d36ac71011
--- /dev/null
+++ b/result/per_paper/2510.15821/accuracy_efficiency.csv
@@ -0,0 +1,13 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Chronos-2,Chronos-2,Avg. Win Rate (%),-2,
+accuracy,TimesFM-2.5,Chronos-2,Avg. Win Rate (%),-2.5,
+accuracy,Moirai-2.0,Chronos-2,Avg. Win Rate (%),-2.0,
+accuracy,Toto-1.0,Chronos-2,Avg. Win Rate (%),-1.0,
+accuracy,Chronos-2,Chronos-2,Avg. Win Rate (%),-2,
+accuracy,TimesFM-2.5,Chronos-2,Avg. Win Rate (%),-2.5,
+accuracy,Toto-1.0,Chronos-2,Avg. Win Rate (%),-1.0,
+accuracy,Moirai-2.0,Chronos-2,Avg. Win Rate (%),-2.0,
+accuracy,Chronos-2,Chronos-2,Avg. Win Rate (%),-2,
+accuracy,TimesFM-2.5,Chronos-2,Avg. Win Rate (%),-2.5,
+accuracy,Toto-1.0,Chronos-2,Avg. Win Rate (%),-1.0,
+accuracy,Moirai-2.0,Chronos-2,Avg. Win Rate (%),-2.0,
diff --git a/result/per_paper/2510.15821/accuracy_efficiency_traced.csv b/result/per_paper/2510.15821/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5e97b14999fed6e097c43f2e2b3ed2c9a871b62a
--- /dev/null
+++ b/result/per_paper/2510.15821/accuracy_efficiency_traced.csv
@@ -0,0 +1,13 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Chronos-2,Chronos-2,Avg. Win Rate (%),-2,,4,1,0
+accuracy,TimesFM-2.5,Chronos-2,Avg. Win Rate (%),-2.5,,4,2,0
+accuracy,Moirai-2.0,Chronos-2,Avg. Win Rate (%),-2.0,,4,4,0
+accuracy,Toto-1.0,Chronos-2,Avg. Win Rate (%),-1.0,,4,5,0
+accuracy,Chronos-2,Chronos-2,Avg. Win Rate (%),-2,,5,1,0
+accuracy,TimesFM-2.5,Chronos-2,Avg. Win Rate (%),-2.5,,5,3,0
+accuracy,Toto-1.0,Chronos-2,Avg. Win Rate (%),-1.0,,5,4,0
+accuracy,Moirai-2.0,Chronos-2,Avg. Win Rate (%),-2.0,,5,5,0
+accuracy,Chronos-2,Chronos-2,Avg. Win Rate (%),-2,,6,1,0
+accuracy,TimesFM-2.5,Chronos-2,Avg. Win Rate (%),-2.5,,6,2,0
+accuracy,Toto-1.0,Chronos-2,Avg. Win Rate (%),-1.0,,6,4,0
+accuracy,Moirai-2.0,Chronos-2,Avg. Win Rate (%),-2.0,,6,5,0
diff --git a/result/per_paper/2510.15821/components_architecture.csv b/result/per_paper/2510.15821/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e6df64a816d90b8489e6d72b52eab0b5951b1658
--- /dev/null
+++ b/result/per_paper/2510.15821/components_architecture.csv
@@ -0,0 +1,8 @@
+component,what_it_is,provenance,citation,evidence
+Group Attention Mechanism,"A mechanism enabling information exchange within groups of time series, supporting univariate, multivariate, and covariate-informed forecasting through in-context learning (ICL). Groups can represent related series, multivariate variates, or targets with covariates.",proposed_here,,Chronos-2 employs a group attention mechanism that facilitates in-context learning (ICL) through efficient information sharing across multiple time series within a group.
+Residual Network for Patch Embeddings,A residual network that maps non-overlapping time series patches (after normalization) into high-dimensional embeddings for the transformer stack.,proposed_here,,The resulting sequences are split into non-overlapping patches and mapped to high-dimensional embeddings via a residual network.
+Transformer Stack with Alternating Time and Group Attention Layers,A core transformer architecture alternating between time attention (aggregates within a single time series) and group attention (aggregates across all series in a group at each patch index).,proposed_here,,The core transformer stack operates on these patch embeddings and produces multi-patch quantile outputs... Each transformer block alternates between time and group attention layers.
+Robust Normalization,A preprocessing step normalizing input time series (targets and covariates) using a robust scaling scheme.,proposed_here,,Input time series (targets and covariates) are first normalized using a robust scaling scheme.
+Time Index and Mask Meta Features,Meta features added to normalized time series to provide temporal context and mask information for future patches during inference.,proposed_here,,"After normalization, time index and mask meta features are added to the input sequences."
+Patch Splitting,A method to divide normalized time series into non-overlapping patches for processing by the residual network and transformer stack.,proposed_here,,The resulting sequences are split into non-overlapping patches and mapped to high-dimensional embeddings.
+Multi-Patch Quantile Output,The final output layer generating quantile predictions for masked future patches across multiple time series.,proposed_here,,The core transformer stack... produces multi-patch quantile outputs corresponding to the masked future patches provided as input.
diff --git a/result/per_paper/2510.15821/computational.csv b/result/per_paper/2510.15821/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..aecfa69fc4b2ec8642345448a9bd7ba262763076
--- /dev/null
+++ b/result/per_paper/2510.15821/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA A10G,,stated,runs on a single mid-range GPU (NVIDIA A10G)
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,300,time series per second,stated,throughput of 300 time series per second
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,"2048, 8192",time steps,stated,"pretrained with 2048, post-trained with 8192"
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2510.16014/accuracy_efficiency.csv b/result/per_paper/2510.16014/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e4d1f61e8dec65087d3add82adcb750394b4c3c7
--- /dev/null
+++ b/result/per_paper/2510.16014/accuracy_efficiency.csv
@@ -0,0 +1,7 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,+STAR,STAR,ACC,0.822,
+accuracy,UniTS,STAR,ACC,0.747,
+accuracy,+STAR,STAR,ACC,0.816,
+accuracy,+STAR,STAR,ACC,0.996,
+accuracy,UniTS,STAR,ACC,0.977,
+accuracy,+STAR,STAR,ACC,0.995,
diff --git a/result/per_paper/2510.16014/accuracy_efficiency_traced.csv b/result/per_paper/2510.16014/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..bff0bd40cead284bf5190f057151d20e7579e6fa
--- /dev/null
+++ b/result/per_paper/2510.16014/accuracy_efficiency_traced.csv
@@ -0,0 +1,7 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,+STAR,STAR,ACC,0.822,,1,3,1
+accuracy,UniTS,STAR,ACC,0.747,,1,4,1
+accuracy,+STAR,STAR,ACC,0.816,,1,5,1
+accuracy,+STAR,STAR,ACC,0.996,,1,7,1
+accuracy,UniTS,STAR,ACC,0.977,,1,8,1
+accuracy,+STAR,STAR,ACC,0.995,,1,9,1
diff --git a/result/per_paper/2510.16014/components_architecture.csv b/result/per_paper/2510.16014/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..74393980bfb57bc8e5d1b7f5ca08a61dc4e0d9d2
--- /dev/null
+++ b/result/per_paper/2510.16014/components_architecture.csv
@@ -0,0 +1,4 @@
+component,what_it_is,provenance,citation,evidence
+Identity-guided State Encoder,"A module that captures complex categorical semantics of state variables through a learnable State Memory, encoding state information guided by both Variable Identity and State Identity.",proposed_here,,"We designed an Identity-guided State Encoder, which effectively captures the complex categorical semantics of state variables through a learnable State Memory."
+Conditional Bottleneck Adapter,"A module that dynamically generates low-rank adaptation parameters conditioned on the current state, flexibly injecting the influence of state variables into the backbone model.",proposed_here,,"We propose a Conditional Bottleneck Adapter, which dynamically generates low-rank adaptation parameters conditioned on the current state, thereby flexibly injecting the influence of state variables into the backbone model."
+Numeral-State Matching Module,A module designed to more effectively detect anomalies inherent to the state variables themselves by aligning numerical and state information.,proposed_here,,We also introduce a Numeral-State Matching module to more effectively detect anomalies inherent to the state variables themselves.
diff --git a/result/per_paper/2510.16014/computational.csv b/result/per_paper/2510.16014/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb201aa1856e288f74fe03771fc33d4e8d88f31a
--- /dev/null
+++ b/result/per_paper/2510.16014/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,
+num_devices,,,not_reported,
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2511.11698/accuracy_efficiency.csv b/result/per_paper/2511.11698/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f6de9b9a111e4aec19dba8acbbc4b92c1c2582ac
--- /dev/null
+++ b/result/per_paper/2511.11698/accuracy_efficiency.csv
@@ -0,0 +1,9 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Moirai 1.0 small,v0,MASE,1.0,
+accuracy,v0,v0,MASE,0,
+accuracy,v1,v0,MASE,1,
+accuracy,v2,v0,MASE,2,
+accuracy,v3,v0,MASE,3,
+accuracy,v4,v0,MASE,4,
+accuracy,v5,v0,MASE,5,
+accuracy,Moirai 2.0,v0,MASE,2.0,
diff --git a/result/per_paper/2511.11698/accuracy_efficiency_traced.csv b/result/per_paper/2511.11698/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..7d330a18b8ddf5bfd93762018e2d5dad5343a27a
--- /dev/null
+++ b/result/per_paper/2511.11698/accuracy_efficiency_traced.csv
@@ -0,0 +1,9 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Moirai 1.0 small,v0,MASE,1.0,,0,1,0
+accuracy,v0,v0,MASE,0,,0,2,0
+accuracy,v1,v0,MASE,1,,0,3,0
+accuracy,v2,v0,MASE,2,,0,4,0
+accuracy,v3,v0,MASE,3,,0,5,0
+accuracy,v4,v0,MASE,4,,0,6,0
+accuracy,v5,v0,MASE,5,,0,7,0
+accuracy,Moirai 2.0,v0,MASE,2.0,,0,8,0
diff --git a/result/per_paper/2511.11698/components_architecture.csv b/result/per_paper/2511.11698/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..3b3c16ea34021f357763f6a9524c9775cdd4cd9c
--- /dev/null
+++ b/result/per_paper/2511.11698/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Decoder-only Architecture,A simplified autoregressive model structure that replaces the masked-encoder design used in Moirai 1.0.,proposed_here,,"replaces masked-encoder training, multi-patch inputs, and mixture-distribution outputs with a simpler decoder-only architecture"
+Quantile Forecasting,A probabilistic forecasting method that predicts quantiles of the target distribution instead of point estimates.,proposed_here,,"the model adopts quantile forecasting and multi-token prediction, improving both probabilistic accuracy and inference efficiency"
+Multi-token Prediction,"A strategy to predict multiple future time steps in a single forward pass, enhancing inference efficiency.",proposed_here,,"the model adopts quantile forecasting and multi-token prediction, improving both probabilistic accuracy and inference efficiency"
+Single Patch Input,A design choice that processes time series data as a single contiguous segment (patch) instead of multiple patches.,proposed_here,,"replaces masked-encoder training, multi-patch inputs, and mixture-distribution outputs with a simpler decoder-only architecture, single patch, and quantile loss"
+Quantile Loss Function,A refined loss function tailored for quantile-based predictions to improve probabilistic forecasting accuracy.,proposed_here,,adoption of quantile-based predictions with a refined loss function
+Recursive Multi-quantile Decoding,A decoding strategy that iteratively generates quantile predictions for multiple future time steps.,proposed_here,,Ablation studies isolate these changes—showing that the decoder-only backbone along with recursive multi-quantile decoding contribute most to the gains
diff --git a/result/per_paper/2511.11698/computational.csv b/result/per_paper/2511.11698/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..355a6f54af6e1b008e9bc26831ed940bee8306b9
--- /dev/null
+++ b/result/per_paper/2511.11698/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,H200 GPU,,stated,running on a single H200 GPU
+num_devices,1,,stated,running on a single H200 GPU
+training_cost,,,not_reported,
+training_batch_size,256,,stated,"trained for 100,000 steps with a batch size of 256"
+training_steps_or_epochs,100000,steps,stated,"trained for 100,000 steps"
+precision,bf16,,stated,using bf16 mixed-precision arithmetic
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,M,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2602.12147/accuracy_efficiency.csv b/result/per_paper/2602.12147/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2602.12147/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2602.12147/accuracy_efficiency_traced.csv b/result/per_paper/2602.12147/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2602.12147/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2602.12147/components_architecture.csv b/result/per_paper/2602.12147/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..76e142cbdcfd3cde600eb7a8e9293cde215303c1
--- /dev/null
+++ b/result/per_paper/2602.12147/components_architecture.csv
@@ -0,0 +1,11 @@
+component,what_it_is,provenance,citation,evidence
+Human-in-the-loop Benchmark Construction Pipeline,A pipeline integrating large language models and human expertise to ensure high data integrity and redefine task formulation.,proposed_here,,"Integrating large language models and human expertise, we establish a human-in-the-loop benchmark construction pipeline to ensure high data integrity..."
+Pattern-Level Evaluation Perspective,An evaluation approach that moves beyond dataset-level assessments by analyzing intrinsic temporal properties via structural time series features.,proposed_here,,we propose a novel pattern-level evaluation perspective that moves beyond traditional dataset-level evaluations based on static meta labels. By leveraging structural time series features...
+Structural Time Series Features,Curated features derived from STL decomposition to characterize intrinsic temporal properties of time series variates.,proposed_here,,"Leveraging STL decomposition, we select a curated set of structural and interpretable time series features to characterize the intrinsic pattern of each variate."
+Binary Encoding Scheme,A method to retrieve variates with identical pattern representations for pattern-specific evaluation.,proposed_here,,"Through a binary encoding scheme, variates exhibiting identical pattern representations can be retrieved for a pattern-specific evaluation."
+Multi-Granular Leaderboard,A leaderboard enabling in-depth analysis and visualized inspection of model performance across diverse patterns and tasks.,proposed_here,,We evaluate 12 TSFMs and establish a multi-granular leaderboard to facilitate in-depth analysis and visualized inspection.
+Task-Centric Benchmark (TIME),A next-generation benchmark with 50 fresh datasets and 98 forecasting tasks for strict zero-shot TSFM evaluation.,proposed_here,,"we introduce TIME, a next-generation task-centric benchmark comprising 50 fresh datasets and 98 forecasting tasks, tailored for strict zero-shot TSFM evaluation free from data leakage."
+Zero-Shot TSFM Evaluation Setup,A framework for evaluating time series foundation models without data leakage across heterogeneous datasets.,proposed_here,,tailored for strict zero-shot TSFM evaluation free from data leakage.
+Task Formulation Alignment,A process to align forecasting configurations with real-world operational requirements and variate predictability.,proposed_here,,redefine task formulation by aligning forecasting configurations with real-world operational requirements and variate predictability.
+Legacy-Constrained Data Coverage,A structural limitation in existing benchmarks where data composition is dominated by reused legacy sources.,reused_cited,"Zhou et al., 2021; Wu et al., 2021",Legacy-Constrained Data Coverage. The data composition of modern benchmarks remains heavily constrained by legacy datasets...
+Task-Centric Evaluation Paradigm,"A paradigm shift from dataset-centric to task-centric evaluation, emphasizing zero-shot generalization across forecasting tasks.",reused_cited,"Hewamalage et al., 2023; Bergmeir, 2024; Brigato et al., 2026",TSFMs have reshaped the evaluation landscape... shifting the dominant paradigm from a dataset-centric regime to a task-centric regime...
diff --git a/result/per_paper/2602.12147/computational.csv b/result/per_paper/2602.12147/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4597a3f6b0175521c9dc7ccaca7052258c12530d
--- /dev/null
+++ b/result/per_paper/2602.12147/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA RTX 3090 GPU,,stated,Experiments are run on a single NVIDIA RTX 3090 GPU (24GB VRAM).
+num_devices,1,,stated,Experiments are run on a single NVIDIA RTX 3090 GPU (24GB VRAM).
+training_cost,,,not_reported,
+training_batch_size,,,not_reported,
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2602.13807/accuracy_efficiency.csv b/result/per_paper/2602.13807/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6fe46e0645a4ba6c4812d638acb6c3232123607b
--- /dev/null
+++ b/result/per_paper/2602.13807/accuracy_efficiency.csv
@@ -0,0 +1,9 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Qwen3-0.6B,Qwen3-0.6B,F1,3,
+accuracy,Qwen3-0.6B,Qwen3-1.7B,F1,3,
+accuracy,Qwen3-1.7B,Qwen3-0.6B,F1,3,
+accuracy,Qwen3-1.7B,Qwen3-1.7B,F1,3,
+accuracy,Qwen3-4B,Qwen3-0.6B,F1,3,
+accuracy,Qwen3-4B,Qwen3-1.7B,F1,3,
+accuracy,Qwen3-8B,Qwen3-0.6B,F1,3,
+accuracy,Qwen3-8B,Qwen3-1.7B,F1,3,
diff --git a/result/per_paper/2602.13807/accuracy_efficiency_traced.csv b/result/per_paper/2602.13807/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6ce380a410762e89b7fe347d6c9505e18f51e334
--- /dev/null
+++ b/result/per_paper/2602.13807/accuracy_efficiency_traced.csv
@@ -0,0 +1,9 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Qwen3-0.6B,Qwen3-0.6B,F1,3,,2,2,0
+accuracy,Qwen3-0.6B,Qwen3-1.7B,F1,3,,2,2,0
+accuracy,Qwen3-1.7B,Qwen3-0.6B,F1,3,,2,3,0
+accuracy,Qwen3-1.7B,Qwen3-1.7B,F1,3,,2,3,0
+accuracy,Qwen3-4B,Qwen3-0.6B,F1,3,,2,4,0
+accuracy,Qwen3-4B,Qwen3-1.7B,F1,3,,2,4,0
+accuracy,Qwen3-8B,Qwen3-0.6B,F1,3,,2,5,0
+accuracy,Qwen3-8B,Qwen3-1.7B,F1,3,,2,5,0
diff --git a/result/per_paper/2602.13807/components_architecture.csv b/result/per_paper/2602.13807/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5ae483013f1b0904ff6fac858f12f320b5b2c431
--- /dev/null
+++ b/result/per_paper/2602.13807/components_architecture.csv
@@ -0,0 +1,16 @@
+component,what_it_is,provenance,citation,evidence
+Coarse-to-Fine Workflow,"A structured workflow that incrementally localizes suspicious intervals, constructs diagnostic evidence, and refines anomaly decisions through sequential steps.",proposed_here,,"AnomaMind operates through a coarse-to-fine workflow that first localizes suspicious intervals, then constructs diagnostic evidence through tool interaction, and finally refines anomaly decisions through self-reflection."
+Toolkit Box,"A modular system integrating knowledge memory (visual patterns, domain knowledge) and numerical diagnostics (statistical, value-based, change-based, region-level operators) for evidence-based verification.",proposed_here,,"The workflow is supported by a toolkit box that combines knowledge memory and numerical diagnostics: visual anomaly patterns mined from training data and domain knowledge provide contextual guidance, while statistical, value-based, change-based, and region-level operators provide measurable evidence for verification."
+Knowledge Memory,"A repository storing visual anomaly patterns from training data, domain knowledge, and tool descriptions to guide contextual reasoning.",proposed_here,,"The knowledge memory includes visual anomaly patterns mined from training data, domain knowledge, and tool descriptions."
+Numerical-Evidence Toolkit,"A set of operators (statistical, value-based, change-based, region-level) for quantifiable verification of anomalies.",proposed_here,,"The numerical-evidence toolkit provides statistical, value-based, change-based, and region-level operators for measurable verification."
+Hybrid Inference Mechanism,"A dual-system architecture where general-purpose models handle flexible reasoning and tool orchestration, while a detection-specific policy is optimized for structured outputs.",proposed_here,,"AnomaMind adopts a hybrid inference mechanism: general-purpose models handle tool invocation and self-reflective refinement, while a detection-specific policy is optimized through rule-based rewards."
+General-Purpose Models,"Models responsible for flexible reasoning, tool invocation, and self-reflection during the diagnostic process.",proposed_here,,General-purpose models handle tool invocation and self-reflective refinement.
+Detection-Specific Policy,"A policy optimized with rule-based rewards to ensure valid parsing, F1-score alignment, and false-positive control.",proposed_here,,"A detection-specific policy is optimized through rule-based rewards that encourage valid parsing, F1-score alignment, and false-positive control."
+Rule-Based Rewards,"A reward mechanism for training the detection-specific policy, enforcing constraints on output validity and performance metrics.",proposed_here,,"A detection-specific policy is optimized with rule-based rewards for parsable outputs, F1-score alignment, and false-positive control."
+Visual Anomaly Patterns,Patterns mined from training data to provide contextual guidance for anomaly localization.,proposed_here,,Visual anomaly patterns mined from training data and domain knowledge provide contextual guidance.
+Domain Knowledge,External knowledge integrated into the knowledge memory to inform contextual reasoning during anomaly detection.,proposed_here,,Visual anomaly patterns mined from training data and domain knowledge provide contextual guidance.
+Tool Descriptions,Descriptions of diagnostic tools stored in the knowledge memory to guide tool interaction during inference.,proposed_here,,The knowledge memory includes ... tool descriptions.
+Statistical Operators,Numerical tools for analyzing statistical properties of time series data to verify anomalies.,proposed_here,,"Statistical, value-based, change-based, and region-level operators provide measurable evidence for verification."
+Value-Based Operators,Tools for evaluating absolute values or thresholds in time series data to detect anomalies.,proposed_here,,"Statistical, value-based, change-based, and region-level operators provide measurable evidence for verification."
+Change-Based Operators,Tools for detecting abrupt or gradual changes in time series data indicative of anomalies.,proposed_here,,"Statistical, value-based, change-based, and region-level operators provide measurable evidence for verification."
+Region-Level Operators,Tools for analyzing anomalies at specific temporal regions or segments of the time series.,proposed_here,,"Statistical, value-based, change-based, and region-level operators provide measurable evidence for verification."
diff --git a/result/per_paper/2602.13807/computational.csv b/result/per_paper/2602.13807/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..27a7e83557b4144e888c3beb282298eac3880fc0
--- /dev/null
+++ b/result/per_paper/2602.13807/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA A800 GPU,GPU,stated,trained on four NVIDIA A800 GPUs (80GB)
+num_devices,4,device,stated,trained on four NVIDIA A800 GPUs (80GB)
+training_cost,,,not_reported,
+training_batch_size,16,samples,stated,batch size of 16 on four NVIDIA A800 GPUs
+training_steps_or_epochs,,,not_reported,
+precision,,,not_reported,
+inference_latency,,,not_reported,
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,,,not_reported,
+context_lengths_evaluated,,,not_reported,
+horizon_lengths_evaluated,,,not_reported,
+inference_batch_size,,,not_reported,
diff --git a/result/per_paper/2602.16681/accuracy_efficiency.csv b/result/per_paper/2602.16681/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..328cf9847753c6803016256e362b952b13a1220f
--- /dev/null
+++ b/result/per_paper/2602.16681/accuracy_efficiency.csv
@@ -0,0 +1,116 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,TimeRCD,VETime,Avg Rank,83.28,
+accuracy,DADA†,VETime,Avg Rank,89.37,
+accuracy,TS-Pulse,VETime,Avg Rank,68.76,
+accuracy,MOMENT†,VETime,Avg Rank,87.54,
+accuracy,TimesFM,VETime,Avg Rank,81.88,
+accuracy,Chronos,VETime,Avg Rank,90.12,
+accuracy,Time MOE,VETime,Avg Rank,76.34,
+accuracy,TimeRCD,VETime,Avg Rank,28.44,
+accuracy,DADA†,VETime,Avg Rank,42.50,
+accuracy,TS-Pulse,VETime,Avg Rank,04.10,
+accuracy,MOMENT†,VETime,Avg Rank,33.15,
+accuracy,TimesFM,VETime,Avg Rank,48.95,
+accuracy,Chronos,VETime,Avg Rank,45.45,
+accuracy,Time MOE,VETime,Avg Rank,25.95,
+accuracy,TimeRCD,VETime,Avg Rank,24.22,
+accuracy,DADA†,VETime,Avg Rank,32.76,
+accuracy,TS-Pulse,VETime,Avg Rank,03.54,
+accuracy,MOMENT†,VETime,Avg Rank,30.69,
+accuracy,TimesFM,VETime,Avg Rank,34.28,
+accuracy,Chronos,VETime,Avg Rank,32.69,
+accuracy,Time MOE,VETime,Avg Rank,26.52,
+accuracy,TimeRCD,VETime,Avg Rank,20.23,
+accuracy,DADA†,VETime,Avg Rank,24.97,
+accuracy,TS-Pulse,VETime,Avg Rank,04.64,
+accuracy,MOMENT†,VETime,Avg Rank,37.35,
+accuracy,TimesFM,VETime,Avg Rank,19.56,
+accuracy,Chronos,VETime,Avg Rank,19.00,
+accuracy,Time MOE,VETime,Avg Rank,16.63,
+accuracy,Grand Total (Zero-Shot),VETime,Avg Rank,25,
+accuracy,TranAD,VETime,Avg Rank,83.19,
+accuracy,USAD,VETime,Avg Rank,71.08,
+accuracy,OmniAnomaly,VETime,Avg Rank,80.32,
+accuracy,LOF,VETime,Avg Rank,81.06,
+accuracy,IForest,VETime,Avg Rank,52.81,
+accuracy,TranAD,VETime,Avg Rank,22.63,
+accuracy,USAD,VETime,Avg Rank,20.99,
+accuracy,OmniAnomaly,VETime,Avg Rank,51.17,
+accuracy,LOF,VETime,Avg Rank,27.97,
+accuracy,IForest,VETime,Avg Rank,07.64,
+accuracy,TranAD,VETime,Avg Rank,34.85,
+accuracy,USAD,VETime,Avg Rank,30.66,
+accuracy,OmniAnomaly,VETime,Avg Rank,47.05,
+accuracy,LOF,VETime,Avg Rank,30.28,
+accuracy,IForest,VETime,Avg Rank,08.37,
+accuracy,TranAD,VETime,Avg Rank,21.61,
+accuracy,USAD,VETime,Avg Rank,16.58,
+accuracy,OmniAnomaly,VETime,Avg Rank,25.35,
+accuracy,LOF,VETime,Avg Rank,19.43,
+accuracy,IForest,VETime,Avg Rank,08.59,
+accuracy,TimeRCD,VETime,F1,81.16,
+accuracy,DADA†,VETime,F1,76.57,
+accuracy,TS-Pulse,VETime,F1,70.14,
+accuracy,MOMENT†,VETime,F1,74.55,
+accuracy,Time MOE,VETime,F1,69.85,
+accuracy,TimeRCD,VETime,F1,42.47,
+accuracy,DADA†,VETime,F1,34.58,
+accuracy,TS-Pulse,VETime,F1,23.57,
+accuracy,MOMENT†,VETime,F1,25.97,
+accuracy,Time MOE,VETime,F1,23.92,
+accuracy,TimeRCD,VETime,F1,30.66,
+accuracy,DADA†,VETime,F1,22.13,
+accuracy,TS-Pulse,VETime,F1,12.56,
+accuracy,MOMENT†,VETime,F1,14.43,
+accuracy,Time MOE,VETime,F1,12.85,
+accuracy,TimeRCD,VETime,F1,20.45,
+accuracy,DADA†,VETime,F1,12.74,
+accuracy,TS-Pulse,VETime,F1,07.41,
+accuracy,MOMENT†,VETime,F1,09.32,
+accuracy,Time MOE,VETime,F1,07.82,
+accuracy,VETime Grand Total (Zero-Shot),VETime,F1,15,
+accuracy,TranAD,VETime,F1,79.91,
+accuracy,USAD,VETime,F1,81.86,
+accuracy,OmniAnomaly,VETime,F1,83.15,
+accuracy,LOF,VETime,F1,84.35,
+accuracy,TranAD,VETime,F1,39.42,
+accuracy,USAD,VETime,F1,48.71,
+accuracy,OmniAnomaly,VETime,F1,49.36,
+accuracy,LOF,VETime,F1,38.97,
+accuracy,TranAD,VETime,F1,29.60,
+accuracy,USAD,VETime,F1,38.71,
+accuracy,OmniAnomaly,VETime,F1,39.10,
+accuracy,LOF,VETime,F1,30.65,
+accuracy,TranAD,VETime,F1,14.78,
+accuracy,USAD,VETime,F1,29.95,
+accuracy,OmniAnomaly,VETime,F1,31.57,
+accuracy,LOF,VETime,F1,24.67,
+accuracy,VETime Grand Total (Full-Shot),VETime,F1,7,
+accuracy,VIT4TS,VETime,F1,61.12,
+accuracy,VLM4TS,VETime,F1,70.37,
+accuracy,VisualTimeAnomaly,VETime,F1,62.46,
+accuracy,AnomLLM,VETime,F1,58.53,
+accuracy,VIT4TS,VETime,F1,28.29,
+accuracy,VLM4TS,VETime,F1,30.75,
+accuracy,VisualTimeAnomaly,VETime,F1,41.64,
+accuracy,AnomLLM,VETime,F1,17.76,
+accuracy,VIT4TS,VETime,F1,33.81,
+accuracy,VLM4TS,VETime,F1,34.78,
+accuracy,VisualTimeAnomaly,VETime,F1,31.72,
+accuracy,AnomLLM,VETime,F1,23.70,
+accuracy,VIT4TS,VETime,F1,25.78,
+accuracy,VLM4TS,VETime,F1,27.89,
+accuracy,VisualTimeAnomaly,VETime,F1,26.87,
+accuracy,AnomLLM,VETime,F1,19.50,
+accuracy,VIT4TS,VETime,F1,1.72,
+accuracy,VLM4TS,VETime,F1,6.78,
+accuracy,VisualTimeAnomaly,VETime,F1,2.52,
+accuracy,AnomLLM,VETime,F1,3.59,
+accuracy,TimeRCD,VETime,F1,74.84,
+accuracy,VLM4TS,VETime,F1,67.11,
+accuracy,TimeRCD,VETime,F1,31.51,
+accuracy,VLM4TS,VETime,F1,21.60,
+accuracy,TimeRCD,VETime,F1,28.34,
+accuracy,VLM4TS,VETime,F1,23.15,
+accuracy,TimeRCD,VETime,F1,61.85,
+accuracy,VLM4TS,VETime,F1,30.57,
diff --git a/result/per_paper/2602.16681/accuracy_efficiency_traced.csv b/result/per_paper/2602.16681/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4547c13868801f3f00d5828140c8a0d67a84fafa
--- /dev/null
+++ b/result/per_paper/2602.16681/accuracy_efficiency_traced.csv
@@ -0,0 +1,116 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,TimeRCD,VETime,Avg Rank,83.28,,0,4,1
+accuracy,DADA†,VETime,Avg Rank,89.37,,0,5,1
+accuracy,TS-Pulse,VETime,Avg Rank,68.76,,0,6,1
+accuracy,MOMENT†,VETime,Avg Rank,87.54,,0,7,1
+accuracy,TimesFM,VETime,Avg Rank,81.88,,0,8,1
+accuracy,Chronos,VETime,Avg Rank,90.12,,0,9,1
+accuracy,Time MOE,VETime,Avg Rank,76.34,,0,10,1
+accuracy,TimeRCD,VETime,Avg Rank,28.44,,0,12,1
+accuracy,DADA†,VETime,Avg Rank,42.50,,0,13,1
+accuracy,TS-Pulse,VETime,Avg Rank,04.10,,0,14,1
+accuracy,MOMENT†,VETime,Avg Rank,33.15,,0,15,1
+accuracy,TimesFM,VETime,Avg Rank,48.95,,0,16,1
+accuracy,Chronos,VETime,Avg Rank,45.45,,0,17,1
+accuracy,Time MOE,VETime,Avg Rank,25.95,,0,18,1
+accuracy,TimeRCD,VETime,Avg Rank,24.22,,0,20,1
+accuracy,DADA†,VETime,Avg Rank,32.76,,0,21,1
+accuracy,TS-Pulse,VETime,Avg Rank,03.54,,0,22,1
+accuracy,MOMENT†,VETime,Avg Rank,30.69,,0,23,1
+accuracy,TimesFM,VETime,Avg Rank,34.28,,0,24,1
+accuracy,Chronos,VETime,Avg Rank,32.69,,0,25,1
+accuracy,Time MOE,VETime,Avg Rank,26.52,,0,26,1
+accuracy,TimeRCD,VETime,Avg Rank,20.23,,0,28,1
+accuracy,DADA†,VETime,Avg Rank,24.97,,0,29,1
+accuracy,TS-Pulse,VETime,Avg Rank,04.64,,0,30,1
+accuracy,MOMENT†,VETime,Avg Rank,37.35,,0,31,1
+accuracy,TimesFM,VETime,Avg Rank,19.56,,0,32,1
+accuracy,Chronos,VETime,Avg Rank,19.00,,0,33,1
+accuracy,Time MOE,VETime,Avg Rank,16.63,,0,34,1
+accuracy,Grand Total (Zero-Shot),VETime,Avg Rank,25,,0,35,1
+accuracy,TranAD,VETime,Avg Rank,83.19,,0,38,1
+accuracy,USAD,VETime,Avg Rank,71.08,,0,39,1
+accuracy,OmniAnomaly,VETime,Avg Rank,80.32,,0,40,1
+accuracy,LOF,VETime,Avg Rank,81.06,,0,41,1
+accuracy,IForest,VETime,Avg Rank,52.81,,0,42,1
+accuracy,TranAD,VETime,Avg Rank,22.63,,0,44,1
+accuracy,USAD,VETime,Avg Rank,20.99,,0,45,1
+accuracy,OmniAnomaly,VETime,Avg Rank,51.17,,0,46,1
+accuracy,LOF,VETime,Avg Rank,27.97,,0,47,1
+accuracy,IForest,VETime,Avg Rank,07.64,,0,48,1
+accuracy,TranAD,VETime,Avg Rank,34.85,,0,50,1
+accuracy,USAD,VETime,Avg Rank,30.66,,0,51,1
+accuracy,OmniAnomaly,VETime,Avg Rank,47.05,,0,52,1
+accuracy,LOF,VETime,Avg Rank,30.28,,0,53,1
+accuracy,IForest,VETime,Avg Rank,08.37,,0,54,1
+accuracy,TranAD,VETime,Avg Rank,21.61,,0,56,1
+accuracy,USAD,VETime,Avg Rank,16.58,,0,57,1
+accuracy,OmniAnomaly,VETime,Avg Rank,25.35,,0,58,1
+accuracy,LOF,VETime,Avg Rank,19.43,,0,59,1
+accuracy,IForest,VETime,Avg Rank,08.59,,0,60,1
+accuracy,TimeRCD,VETime,F1,81.16,,1,4,1
+accuracy,DADA†,VETime,F1,76.57,,1,5,1
+accuracy,TS-Pulse,VETime,F1,70.14,,1,6,1
+accuracy,MOMENT†,VETime,F1,74.55,,1,7,1
+accuracy,Time MOE,VETime,F1,69.85,,1,8,1
+accuracy,TimeRCD,VETime,F1,42.47,,1,10,1
+accuracy,DADA†,VETime,F1,34.58,,1,11,1
+accuracy,TS-Pulse,VETime,F1,23.57,,1,12,1
+accuracy,MOMENT†,VETime,F1,25.97,,1,13,1
+accuracy,Time MOE,VETime,F1,23.92,,1,14,1
+accuracy,TimeRCD,VETime,F1,30.66,,1,16,1
+accuracy,DADA†,VETime,F1,22.13,,1,17,1
+accuracy,TS-Pulse,VETime,F1,12.56,,1,18,1
+accuracy,MOMENT†,VETime,F1,14.43,,1,19,1
+accuracy,Time MOE,VETime,F1,12.85,,1,20,1
+accuracy,TimeRCD,VETime,F1,20.45,,1,22,1
+accuracy,DADA†,VETime,F1,12.74,,1,23,1
+accuracy,TS-Pulse,VETime,F1,07.41,,1,24,1
+accuracy,MOMENT†,VETime,F1,09.32,,1,25,1
+accuracy,Time MOE,VETime,F1,07.82,,1,26,1
+accuracy,VETime Grand Total (Zero-Shot),VETime,F1,15,,1,27,1
+accuracy,TranAD,VETime,F1,79.91,,1,30,1
+accuracy,USAD,VETime,F1,81.86,,1,31,1
+accuracy,OmniAnomaly,VETime,F1,83.15,,1,32,1
+accuracy,LOF,VETime,F1,84.35,,1,33,1
+accuracy,TranAD,VETime,F1,39.42,,1,35,1
+accuracy,USAD,VETime,F1,48.71,,1,36,1
+accuracy,OmniAnomaly,VETime,F1,49.36,,1,37,1
+accuracy,LOF,VETime,F1,38.97,,1,38,1
+accuracy,TranAD,VETime,F1,29.60,,1,40,1
+accuracy,USAD,VETime,F1,38.71,,1,41,1
+accuracy,OmniAnomaly,VETime,F1,39.10,,1,42,1
+accuracy,LOF,VETime,F1,30.65,,1,43,1
+accuracy,TranAD,VETime,F1,14.78,,1,45,1
+accuracy,USAD,VETime,F1,29.95,,1,46,1
+accuracy,OmniAnomaly,VETime,F1,31.57,,1,47,1
+accuracy,LOF,VETime,F1,24.67,,1,48,1
+accuracy,VETime Grand Total (Full-Shot),VETime,F1,7,,1,49,1
+accuracy,VIT4TS,VETime,F1,61.12,,2,2,1
+accuracy,VLM4TS,VETime,F1,70.37,,2,3,1
+accuracy,VisualTimeAnomaly,VETime,F1,62.46,,2,4,1
+accuracy,AnomLLM,VETime,F1,58.53,,2,5,1
+accuracy,VIT4TS,VETime,F1,28.29,,2,7,1
+accuracy,VLM4TS,VETime,F1,30.75,,2,8,1
+accuracy,VisualTimeAnomaly,VETime,F1,41.64,,2,9,1
+accuracy,AnomLLM,VETime,F1,17.76,,2,10,1
+accuracy,VIT4TS,VETime,F1,33.81,,2,12,1
+accuracy,VLM4TS,VETime,F1,34.78,,2,13,1
+accuracy,VisualTimeAnomaly,VETime,F1,31.72,,2,14,1
+accuracy,AnomLLM,VETime,F1,23.70,,2,15,1
+accuracy,VIT4TS,VETime,F1,25.78,,2,17,1
+accuracy,VLM4TS,VETime,F1,27.89,,2,18,1
+accuracy,VisualTimeAnomaly,VETime,F1,26.87,,2,19,1
+accuracy,AnomLLM,VETime,F1,19.50,,2,20,1
+accuracy,VIT4TS,VETime,F1,1.72,,2,22,1
+accuracy,VLM4TS,VETime,F1,6.78,,2,23,1
+accuracy,VisualTimeAnomaly,VETime,F1,2.52,,2,24,1
+accuracy,AnomLLM,VETime,F1,3.59,,2,25,1
+accuracy,TimeRCD,VETime,F1,74.84,,4,2,1
+accuracy,VLM4TS,VETime,F1,67.11,,4,3,1
+accuracy,TimeRCD,VETime,F1,31.51,,4,5,1
+accuracy,VLM4TS,VETime,F1,21.60,,4,6,1
+accuracy,TimeRCD,VETime,F1,28.34,,4,8,1
+accuracy,VLM4TS,VETime,F1,23.15,,4,9,1
+accuracy,TimeRCD,VETime,F1,61.85,,4,11,1
+accuracy,VLM4TS,VETime,F1,30.57,,4,12,1
diff --git a/result/per_paper/2602.16681/components_architecture.csv b/result/per_paper/2602.16681/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..87ef5f3391423ccb844eef47ef0258008a146a41
--- /dev/null
+++ b/result/per_paper/2602.16681/components_architecture.csv
@@ -0,0 +1,5 @@
+component,what_it_is,provenance,citation,evidence
+Reversible Image Conversion,A module that constructs information-dense images with discriminative anomaly details by converting time series data into visual representations.,proposed_here,"Yingyuan Yang et al., 2025",VETime introduces a Reversible Image Conversion and a Patch-Level Temporal Alignment module to establish a shared visual-temporal timeline...
+Patch-Level Temporal Alignment,"A module that enhances 2D visual representations from pre-trained ViT with 1D temporal ordering, enabling fine-grained cross-modal interaction.",proposed_here,"Yingyuan Yang et al., 2025","Complementing this, we propose the Patch-Level Temporal Alignment module, which enhances 2D visual representations from the pre-trained ViT with a 1D temporal ordering..."
+Anomaly Window Contrastive Learning,A mechanism that incorporates intra- and inter-window contrast between multi-modal features to achieve comprehensive anomaly identification.,proposed_here,"Yingyuan Yang et al., 2025","Considering the distinct perceptual characteristics of visual and temporal modalities regarding anomalies, we propose a specialized Anomaly Window Contrastive Learning..."
+Task-Adaptive Multi-Modal Fusion,A module that dynamically and adaptively weights fused features from aligned modalities to ensure precise downstream anomaly detection.,proposed_here,"Yingyuan Yang et al., 2025",Task-Adaptive Multi-Modal Fusion module equipped with a reconstruction head is applied to facilitate robust multi-modal fusion...
diff --git a/result/per_paper/2602.16681/computational.csv b/result/per_paper/2602.16681/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b48ebdbbc41ecbe502d5d0f627276707d837539f
--- /dev/null
+++ b/result/per_paper/2602.16681/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No hardware specifications (e.g., GPU/TPU types) are mentioned in the text or tables."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not specified."
+training_cost,,,not_reported,No monetary or resource-based training cost is reported.
+training_batch_size,,,not_reported,Batch size during training is not mentioned in the text or tables.
+training_steps_or_epochs,,,not_reported,The number of training steps or epochs is not provided.
+precision,,,not_reported,"Computational precision (e.g., 32-bit vs. 16-bit) is not specified."
+inference_latency,0.03,seconds,stated,"Per-series runtime for VETime is reported as 0.04, 0.02, 0.04, 0.02 s for NAB, YAHOO, SMAP, MSL (average ≈ 0.03 s)."
+inference_throughput,,,not_reported,Throughput (samples/second) is not explicitly reported.
+peak_memory,,,not_reported,Peak memory usage during training or inference is not mentioned.
+flops_or_macs,,,not_reported,FLOPs or MACs are not provided in the text or tables.
+num_inference_samples,,,not_reported,The number of inference samples tested is not specified.
+params,,,not_reported,Model parameter count is not reported.
+context_lengths_evaluated,,,not_reported,Specific context lengths evaluated are not mentioned.
+horizon_lengths_evaluated,,,not_reported,Horizon lengths evaluated are not specified.
+inference_batch_size,,,not_reported,Inference batch size is not explicitly stated.
diff --git a/result/per_paper/2602.17634/accuracy_efficiency.csv b/result/per_paper/2602.17634/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ce9f0e9792650401d3266ba8ac530b2c67273fa9
--- /dev/null
+++ b/result/per_paper/2602.17634/accuracy_efficiency.csv
@@ -0,0 +1,126 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,loop_seattle/5T/short,Reverso,MASE,0.5531,
+accuracy,loop_seattle/5T/short,Reverso,MASE,0.8112,
+accuracy,loop_seattle/5T/medium,Reverso,MASE,0.8008,
+accuracy,loop_seattle/5T/medium,Reverso,MASE,1.2476,
+accuracy,loop_seattle/5T/long,Reverso,MASE,0.8880,
+accuracy,loop_seattle/5T/long,Reverso,MASE,1.3336,
+accuracy,loop_seattle/D/short,Reverso,MASE,0.8818,
+accuracy,loop_seattle/D/short,Reverso,MASE,1.4936,
+accuracy,loop_seattle/H/short,Reverso,MASE,0.8113,
+accuracy,loop_seattle/H/short,Reverso,MASE,0.7479,
+accuracy,loop_seattle/H/medium,Reverso,MASE,0.9009,
+accuracy,loop_seattle/H/medium,Reverso,MASE,0.8903,
+accuracy,loop_seattle/H/long,Reverso,MASE,0.8668,
+accuracy,loop_seattle/H/long,Reverso,MASE,0.9261,
+accuracy,m_dense/D/short,Reverso,MASE,0.7038,
+accuracy,m_dense/D/short,Reverso,MASE,1.3449,
+accuracy,m_dense/H/short,Reverso,MASE,0.8146,
+accuracy,m_dense/H/short,Reverso,MASE,0.7140,
+accuracy,m_dense/H/medium,Reverso,MASE,0.6703,
+accuracy,m_dense/H/medium,Reverso,MASE,0.9900,
+accuracy,m_dense/H/long,Reverso,MASE,0.6777,
+accuracy,m_dense/H/long,Reverso,MASE,0.9873,
+accuracy,sz_taxi/15T/short,Reverso,MASE,0.5441,
+accuracy,sz_taxi/15T/short,Reverso,MASE,0.7538,
+accuracy,sz_taxi/15T/medium,Reverso,MASE,0.5420,
+accuracy,sz_taxi/15T/medium,Reverso,MASE,0.7456,
+accuracy,sz_taxi/15T/long,Reverso,MASE,0.5148,
+accuracy,sz_taxi/15T/long,Reverso,MASE,0.7469,
+accuracy,sz_taxi/H/short,Reverso,MASE,0.5698,
+accuracy,sz_taxi/H/short,Reverso,MASE,0.7863,
+accuracy,bitbrains_fast_storage/5T/short,Reverso,MASE,0.6703,
+accuracy,bitbrains_fast_storage/5T/short,Reverso,MASE,0.2907,
+accuracy,bitbrains_fast_storage/5T/medium,Reverso,MASE,0.9883,
+accuracy,bitbrains_fast_storage/5T/medium,Reverso,MASE,0.6179,
+accuracy,bitbrains_fast_storage/5T/long,Reverso,MASE,0.8828,
+accuracy,bitbrains_fast_storage/5T/long,Reverso,MASE,0.6427,
+accuracy,bitbrains_fast_storage/H/short,Reverso,MASE,1.0029,
+accuracy,bitbrains_fast_storage/H/short,Reverso,MASE,1.3003,
+accuracy,bitbrains_rnd/5T/short,Reverso,MASE,1.6286,
+accuracy,bitbrains_rnd/5T/short,Reverso,MASE,0.5126,
+accuracy,bitbrains_rnd/5T/medium,Reverso,MASE,4.3766,
+accuracy,bitbrains_rnd/5T/medium,Reverso,MASE,0.7812,
+accuracy,bitbrains_rnd/5T/long,Reverso,MASE,3.3422,
+accuracy,bitbrains_rnd/5T/long,Reverso,MASE,1.0075,
+accuracy,bitbrains_rnd/H/short,Reverso,MASE,5.8132,
+accuracy,bitbrains_rnd/H/short,Reverso,MASE,1.2126,
+accuracy,bizitobs_application/10S/short,Reverso,MASE,1.0614,
+accuracy,bizitobs_application/10S/short,Reverso,MASE,0.9513,
+accuracy,bizitobs_application/10S/medium,Reverso,MASE,1.5193,
+accuracy,bizitobs_application/10S/medium,Reverso,MASE,1.0454,
+accuracy,bizitobs_application/10S/long,Reverso,MASE,3.2373,
+accuracy,bizitobs_application/10S/long,Reverso,MASE,1.0415,
+accuracy,bizitobs_12c/5T/short,Reverso,MASE,0.2825,
+accuracy,bizitobs_12c/5T/short,Reverso,MASE,3.3280,
+accuracy,bizitobs_12c/5T/medium,Reverso,MASE,0.4683,
+accuracy,bizitobs_12c/5T/medium,Reverso,MASE,0.7820,
+accuracy,bizitobs_12c/5T/long,Reverso,MASE,0.4891,
+accuracy,bizitobs_12c/5T/long,Reverso,MASE,0.9472,
+accuracy,bizitobs_12c/H/short,Reverso,MASE,0.4383,
+accuracy,bizitobs_12c/H/short,Reverso,MASE,1.2159,
+accuracy,bizitobs_12c/H/medium,Reverso,MASE,0.4881,
+accuracy,bizitobs_12c/H/medium,Reverso,MASE,2.0522,
+accuracy,bizitobs_12c/H/long,Reverso,MASE,0.5473,
+accuracy,bizitobs_12c/H/long,Reverso,MASE,3.4251,
+accuracy,bizitobs_service/10S/short,Reverso,MASE,0.7725,
+accuracy,bizitobs_service/10S/short,Reverso,MASE,0.6936,
+accuracy,bizitobs_service/10S/medium,Reverso,MASE,0.9605,
+accuracy,bizitobs_service/10S/medium,Reverso,MASE,2.8601,
+accuracy,bizitobs_service/10S/long,Reverso,MASE,1.3776,
+accuracy,bizitobs_service/10S/long,Reverso,MASE,0.7681,
+accuracy,car_parts/M/short,Reverso,MASE,0.8652,
+accuracy,car_parts/M/short,Reverso,MASE,1.2569,
+accuracy,covid_deaths/D/short,Reverso,MASE,34.3613,
+accuracy,covid_deaths/D/short,Reverso,MASE,1.1669,
+accuracy,electricity/15T/short,Reverso,MASE,1.0297,
+accuracy,electricity/15T/short,Reverso,MASE,0.8224,
+accuracy,electricity/15T/medium,Reverso,MASE,0.8463,
+accuracy,electricity/15T/medium,Reverso,MASE,0.8593,
+accuracy,electricity/15T/long,Reverso,MASE,0.8917,
+accuracy,electricity/15T/long,Reverso,MASE,1.0168,
+accuracy,electricity/D/short,Reverso,MASE,1.4781,
+accuracy,electricity/D/short,Reverso,MASE,0.8386,
+accuracy,electricity/H/short,Reverso,MASE,0.9649,
+accuracy,electricity/H/short,Reverso,MASE,0.8806,
+accuracy,electricity/H/medium,Reverso,MASE,1.0626,
+accuracy,electricity/H/medium,Reverso,MASE,0.9416,
+accuracy,electricity/H/long,Reverso,MASE,1.1940,
+accuracy,electricity/H/long,Reverso,MASE,1.4687,
+accuracy,electricity/W/short,Reverso,MASE,1.6038,
+accuracy,electricity/W/short,Reverso,MASE,1.3853,
+accuracy,ett1/15T/short,Reverso,MASE,0.6924,
+accuracy,ett1/15T/short,Reverso,MASE,0.3505,
+accuracy,ett1/15T/medium,Reverso,MASE,1.0419,
+accuracy,ett1/15T/medium,Reverso,MASE,0.7934,
+accuracy,ett1/15T/long,Reverso,MASE,1.0498,
+accuracy,ett1/15T/long,Reverso,MASE,1.0961,
+accuracy,ett1/D/short,Reverso,MASE,1.6066,
+accuracy,Params,Reverso,ETTm1,2.6,
+accuracy,Params,Reverso-Small,ETTm1,550,
+accuracy,Params,Reverso-Nano,ETTm1,200,
+accuracy,ETTm1,Reverso,ETTm1,0.367,
+accuracy,ETTm1,Reverso-Small,ETTm1,0.376,
+accuracy,ETTm1,Reverso-Nano,ETTm1,0.382,
+accuracy,ETTm2,Reverso,ETTm1,0.304,
+accuracy,ETTm2,Reverso-Small,ETTm1,0.309,
+accuracy,ETTm2,Reverso-Nano,ETTm1,0.311,
+accuracy,ETTh1,Reverso,ETTm1,0.404,
+accuracy,ETTh1,Reverso-Small,ETTm1,0.404,
+accuracy,ETTh1,Reverso-Nano,ETTm1,0.416,
+accuracy,ETTh2,Reverso,ETTm1,0.365,
+accuracy,ETTh2,Reverso-Small,ETTm1,0.370,
+accuracy,ETTh2,Reverso-Nano,ETTm1,0.384,
+accuracy,Electricity,Reverso,ETTm1,0.238,
+accuracy,Electricity,Reverso-Small,ETTm1,0.241,
+accuracy,Electricity,Reverso-Nano,ETTm1,0.249,
+accuracy,Weather,Reverso,ETTm1,0.253,
+accuracy,Weather,Reverso-Small,ETTm1,0.252,
+accuracy,Weather,Reverso-Nano,ETTm1,0.257,
+accuracy,Avg,Reverso,ETTm1,0.322,
+accuracy,Avg,Reverso-Small,ETTm1,0.325,
+accuracy,Avg,Reverso-Nano,ETTm1,0.333,
+accuracy,Avg Rank,Reverso,ETTm1,2.50,
+accuracy,Avg Rank,Reverso-Small,ETTm1,3.83,
+accuracy,Avg Rank,Reverso-Nano,ETTm1,5.17,
+accuracy,TimesFM-2.5,Reverso,MASE,-2.5,
diff --git a/result/per_paper/2602.17634/accuracy_efficiency_traced.csv b/result/per_paper/2602.17634/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..f93993997f1647136a34dc992eeabafe9568435f
--- /dev/null
+++ b/result/per_paper/2602.17634/accuracy_efficiency_traced.csv
@@ -0,0 +1,126 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,loop_seattle/5T/short,Reverso,MASE,0.5531,,1,1,3
+accuracy,loop_seattle/5T/short,Reverso,MASE,0.8112,,1,1,7
+accuracy,loop_seattle/5T/medium,Reverso,MASE,0.8008,,1,2,3
+accuracy,loop_seattle/5T/medium,Reverso,MASE,1.2476,,1,2,7
+accuracy,loop_seattle/5T/long,Reverso,MASE,0.8880,,1,3,3
+accuracy,loop_seattle/5T/long,Reverso,MASE,1.3336,,1,3,7
+accuracy,loop_seattle/D/short,Reverso,MASE,0.8818,,1,4,3
+accuracy,loop_seattle/D/short,Reverso,MASE,1.4936,,1,4,7
+accuracy,loop_seattle/H/short,Reverso,MASE,0.8113,,1,5,3
+accuracy,loop_seattle/H/short,Reverso,MASE,0.7479,,1,5,7
+accuracy,loop_seattle/H/medium,Reverso,MASE,0.9009,,1,6,3
+accuracy,loop_seattle/H/medium,Reverso,MASE,0.8903,,1,6,7
+accuracy,loop_seattle/H/long,Reverso,MASE,0.8668,,1,7,3
+accuracy,loop_seattle/H/long,Reverso,MASE,0.9261,,1,7,7
+accuracy,m_dense/D/short,Reverso,MASE,0.7038,,1,8,3
+accuracy,m_dense/D/short,Reverso,MASE,1.3449,,1,8,7
+accuracy,m_dense/H/short,Reverso,MASE,0.8146,,1,9,3
+accuracy,m_dense/H/short,Reverso,MASE,0.7140,,1,9,7
+accuracy,m_dense/H/medium,Reverso,MASE,0.6703,,1,10,3
+accuracy,m_dense/H/medium,Reverso,MASE,0.9900,,1,10,7
+accuracy,m_dense/H/long,Reverso,MASE,0.6777,,1,11,3
+accuracy,m_dense/H/long,Reverso,MASE,0.9873,,1,11,7
+accuracy,sz_taxi/15T/short,Reverso,MASE,0.5441,,1,12,3
+accuracy,sz_taxi/15T/short,Reverso,MASE,0.7538,,1,12,7
+accuracy,sz_taxi/15T/medium,Reverso,MASE,0.5420,,1,13,3
+accuracy,sz_taxi/15T/medium,Reverso,MASE,0.7456,,1,13,7
+accuracy,sz_taxi/15T/long,Reverso,MASE,0.5148,,1,14,3
+accuracy,sz_taxi/15T/long,Reverso,MASE,0.7469,,1,14,7
+accuracy,sz_taxi/H/short,Reverso,MASE,0.5698,,1,15,3
+accuracy,sz_taxi/H/short,Reverso,MASE,0.7863,,1,15,7
+accuracy,bitbrains_fast_storage/5T/short,Reverso,MASE,0.6703,,1,16,3
+accuracy,bitbrains_fast_storage/5T/short,Reverso,MASE,0.2907,,1,16,7
+accuracy,bitbrains_fast_storage/5T/medium,Reverso,MASE,0.9883,,1,17,3
+accuracy,bitbrains_fast_storage/5T/medium,Reverso,MASE,0.6179,,1,17,7
+accuracy,bitbrains_fast_storage/5T/long,Reverso,MASE,0.8828,,1,18,3
+accuracy,bitbrains_fast_storage/5T/long,Reverso,MASE,0.6427,,1,18,7
+accuracy,bitbrains_fast_storage/H/short,Reverso,MASE,1.0029,,1,19,3
+accuracy,bitbrains_fast_storage/H/short,Reverso,MASE,1.3003,,1,19,7
+accuracy,bitbrains_rnd/5T/short,Reverso,MASE,1.6286,,1,20,3
+accuracy,bitbrains_rnd/5T/short,Reverso,MASE,0.5126,,1,20,7
+accuracy,bitbrains_rnd/5T/medium,Reverso,MASE,4.3766,,1,21,3
+accuracy,bitbrains_rnd/5T/medium,Reverso,MASE,0.7812,,1,21,7
+accuracy,bitbrains_rnd/5T/long,Reverso,MASE,3.3422,,1,22,3
+accuracy,bitbrains_rnd/5T/long,Reverso,MASE,1.0075,,1,22,7
+accuracy,bitbrains_rnd/H/short,Reverso,MASE,5.8132,,1,23,3
+accuracy,bitbrains_rnd/H/short,Reverso,MASE,1.2126,,1,23,7
+accuracy,bizitobs_application/10S/short,Reverso,MASE,1.0614,,1,24,3
+accuracy,bizitobs_application/10S/short,Reverso,MASE,0.9513,,1,24,7
+accuracy,bizitobs_application/10S/medium,Reverso,MASE,1.5193,,1,25,3
+accuracy,bizitobs_application/10S/medium,Reverso,MASE,1.0454,,1,25,7
+accuracy,bizitobs_application/10S/long,Reverso,MASE,3.2373,,1,26,3
+accuracy,bizitobs_application/10S/long,Reverso,MASE,1.0415,,1,26,7
+accuracy,bizitobs_12c/5T/short,Reverso,MASE,0.2825,,1,27,3
+accuracy,bizitobs_12c/5T/short,Reverso,MASE,3.3280,,1,27,7
+accuracy,bizitobs_12c/5T/medium,Reverso,MASE,0.4683,,1,28,3
+accuracy,bizitobs_12c/5T/medium,Reverso,MASE,0.7820,,1,28,7
+accuracy,bizitobs_12c/5T/long,Reverso,MASE,0.4891,,1,29,3
+accuracy,bizitobs_12c/5T/long,Reverso,MASE,0.9472,,1,29,7
+accuracy,bizitobs_12c/H/short,Reverso,MASE,0.4383,,1,30,3
+accuracy,bizitobs_12c/H/short,Reverso,MASE,1.2159,,1,30,7
+accuracy,bizitobs_12c/H/medium,Reverso,MASE,0.4881,,1,31,3
+accuracy,bizitobs_12c/H/medium,Reverso,MASE,2.0522,,1,31,7
+accuracy,bizitobs_12c/H/long,Reverso,MASE,0.5473,,1,32,3
+accuracy,bizitobs_12c/H/long,Reverso,MASE,3.4251,,1,32,7
+accuracy,bizitobs_service/10S/short,Reverso,MASE,0.7725,,1,33,3
+accuracy,bizitobs_service/10S/short,Reverso,MASE,0.6936,,1,33,7
+accuracy,bizitobs_service/10S/medium,Reverso,MASE,0.9605,,1,34,3
+accuracy,bizitobs_service/10S/medium,Reverso,MASE,2.8601,,1,34,7
+accuracy,bizitobs_service/10S/long,Reverso,MASE,1.3776,,1,35,3
+accuracy,bizitobs_service/10S/long,Reverso,MASE,0.7681,,1,35,7
+accuracy,car_parts/M/short,Reverso,MASE,0.8652,,1,36,3
+accuracy,car_parts/M/short,Reverso,MASE,1.2569,,1,36,7
+accuracy,covid_deaths/D/short,Reverso,MASE,34.3613,,1,37,3
+accuracy,covid_deaths/D/short,Reverso,MASE,1.1669,,1,37,7
+accuracy,electricity/15T/short,Reverso,MASE,1.0297,,1,38,3
+accuracy,electricity/15T/short,Reverso,MASE,0.8224,,1,38,7
+accuracy,electricity/15T/medium,Reverso,MASE,0.8463,,1,39,3
+accuracy,electricity/15T/medium,Reverso,MASE,0.8593,,1,39,7
+accuracy,electricity/15T/long,Reverso,MASE,0.8917,,1,40,3
+accuracy,electricity/15T/long,Reverso,MASE,1.0168,,1,40,7
+accuracy,electricity/D/short,Reverso,MASE,1.4781,,1,41,3
+accuracy,electricity/D/short,Reverso,MASE,0.8386,,1,41,7
+accuracy,electricity/H/short,Reverso,MASE,0.9649,,1,42,3
+accuracy,electricity/H/short,Reverso,MASE,0.8806,,1,42,7
+accuracy,electricity/H/medium,Reverso,MASE,1.0626,,1,43,3
+accuracy,electricity/H/medium,Reverso,MASE,0.9416,,1,43,7
+accuracy,electricity/H/long,Reverso,MASE,1.1940,,1,44,3
+accuracy,electricity/H/long,Reverso,MASE,1.4687,,1,44,7
+accuracy,electricity/W/short,Reverso,MASE,1.6038,,1,45,3
+accuracy,electricity/W/short,Reverso,MASE,1.3853,,1,45,7
+accuracy,ett1/15T/short,Reverso,MASE,0.6924,,1,46,3
+accuracy,ett1/15T/short,Reverso,MASE,0.3505,,1,46,7
+accuracy,ett1/15T/medium,Reverso,MASE,1.0419,,1,47,3
+accuracy,ett1/15T/medium,Reverso,MASE,0.7934,,1,47,7
+accuracy,ett1/15T/long,Reverso,MASE,1.0498,,1,48,3
+accuracy,ett1/15T/long,Reverso,MASE,1.0961,,1,48,7
+accuracy,ett1/D/short,Reverso,MASE,1.6066,,1,49,3
+accuracy,Params,Reverso,ETTm1,2.6,,2,1,1
+accuracy,Params,Reverso-Small,ETTm1,550,,2,1,2
+accuracy,Params,Reverso-Nano,ETTm1,200,,2,1,3
+accuracy,ETTm1,Reverso,ETTm1,0.367,,2,2,1
+accuracy,ETTm1,Reverso-Small,ETTm1,0.376,,2,2,2
+accuracy,ETTm1,Reverso-Nano,ETTm1,0.382,,2,2,3
+accuracy,ETTm2,Reverso,ETTm1,0.304,,2,3,1
+accuracy,ETTm2,Reverso-Small,ETTm1,0.309,,2,3,2
+accuracy,ETTm2,Reverso-Nano,ETTm1,0.311,,2,3,3
+accuracy,ETTh1,Reverso,ETTm1,0.404,,2,4,1
+accuracy,ETTh1,Reverso-Small,ETTm1,0.404,,2,4,2
+accuracy,ETTh1,Reverso-Nano,ETTm1,0.416,,2,4,3
+accuracy,ETTh2,Reverso,ETTm1,0.365,,2,5,1
+accuracy,ETTh2,Reverso-Small,ETTm1,0.370,,2,5,2
+accuracy,ETTh2,Reverso-Nano,ETTm1,0.384,,2,5,3
+accuracy,Electricity,Reverso,ETTm1,0.238,,2,6,1
+accuracy,Electricity,Reverso-Small,ETTm1,0.241,,2,6,2
+accuracy,Electricity,Reverso-Nano,ETTm1,0.249,,2,6,3
+accuracy,Weather,Reverso,ETTm1,0.253,,2,7,1
+accuracy,Weather,Reverso-Small,ETTm1,0.252,,2,7,2
+accuracy,Weather,Reverso-Nano,ETTm1,0.257,,2,7,3
+accuracy,Avg,Reverso,ETTm1,0.322,,2,8,1
+accuracy,Avg,Reverso-Small,ETTm1,0.325,,2,8,2
+accuracy,Avg,Reverso-Nano,ETTm1,0.333,,2,8,3
+accuracy,Avg Rank,Reverso,ETTm1,2.50,,2,9,1
+accuracy,Avg Rank,Reverso-Small,ETTm1,3.83,,2,9,2
+accuracy,Avg Rank,Reverso-Nano,ETTm1,5.17,,2,9,3
+accuracy,TimesFM-2.5,Reverso,MASE,-2.5,,6,3,0
diff --git a/result/per_paper/2602.17634/components_architecture.csv b/result/per_paper/2602.17634/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..cddea120c7908cb27461ee4b07e6ede6904d3d76
--- /dev/null
+++ b/result/per_paper/2602.17634/components_architecture.csv
@@ -0,0 +1,8 @@
+component,what_it_is,provenance,citation,evidence
+Hybrid Model Architecture,A model structure that interleaves long convolution layers and linear RNN layers (specifically DeltaNet layers) to achieve efficient time series forecasting.,proposed_here,,We show that small hybrid models that interleave long convolution and linear RNN layers (in particular DeltaNet layers) can match the performance of larger transformer-based models while being more than a hundred times smaller.
+DeltaNet Layers,"Modern linear RNN layers designed for efficient time series modeling, incorporating mechanisms for temporal dependency capture.",reused_cited,"Schlag et al., 2021; Yang et al., 2024b","small hybrid models that interleave long convolution layers (Fu et al., 2023b) and modern linear RNN layers (in particular DeltaNet layers (Schlag et al., 2021; Yang et al., 2024b))"
+Long Convolution Layers,Convolutional layers with extended receptive fields to capture long-term temporal patterns in time series data.,reused_cited,"Fu et al., 2023b","small hybrid models that interleave long convolution layers (Fu et al., 2023b) and modern linear RNN layers"
+Data Augmentation Strategies,"Techniques to enhance training data diversity, including temporal shifting, scaling, and noise injection, to improve model robustness and generalization.",proposed_here,,We also describe several data augmentation and inference strategies that further improve performance.
+Inference-Time Strategies,"Optimized methods for efficient prediction during deployment, such as context-length adaptation and selective attention mechanisms.",proposed_here,,We also describe several data augmentation and inference strategies that further improve performance.
+Zero-shot Forecasting Capability,"The ability to predict future values of unseen time series domains using in-context learning, without task-specific training.",reused_cited,"Garza et al., 2023; Ansari et al., 2024; Das et al., 2024; Liu et al., 2024; Woo et al., 2024; Liu et al., 2025b;c; Graf et al., 2025; Auer et al., 2025; Moroshan et al., 2025","decoder-based TSFMs are their ability to perform zero-shot forecasting via in-context learning, i.e., predicting the future given any historical time series data given as context."
+Efficient Parameter Scaling,"A design principle to reduce model size while maintaining performance, resulting in models with 0.2M to 2.6M parameters.",proposed_here,,We train a family of TSFMs (dubbed Reverso) from 0.2M to 2.6M parameters that significantly push the performance-efficiency frontier.
diff --git a/result/per_paper/2602.17634/computational.csv b/result/per_paper/2602.17634/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..9edc9b021101c1732d11a174a463425757c32f2c
--- /dev/null
+++ b/result/per_paper/2602.17634/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No information on hardware type (e.g., GPU/TPU) used for training or inference."
+num_devices,,,not_reported,"No mention of the number of devices (e.g., GPUs) used for training or inference."
+training_cost,,,not_reported,"No explicit reporting of training cost (e.g., monetary cost, computational resources)."
+training_batch_size,,,not_reported,No mention of training batch size in the text or tables.
+training_steps_or_epochs,,,not_reported,No information on training steps or epochs provided.
+precision,,,not_reported,"No mention of training or inference precision (e.g., 32-bit, 16-bit)."
+inference_latency,,,not_reported,"No latency metrics (e.g., time per inference) reported."
+inference_throughput,,,not_reported,"No throughput metrics (e.g., samples/second) reported."
+peak_memory,,,not_reported,"No peak memory usage (e.g., GPU/VRAM) reported."
+flops_or_macs,,,not_reported,No FLOPs or MACs computation metrics provided.
+num_inference_samples,,,not_reported,No explicit count of inference samples tested.
+params,2600000,parameters,stated,Reverso has 2.6M parameters (from the 'Model | Parameters | Layers | Dim (d)' table).
+context_lengths_evaluated,,,not_reported,No specific context lengths evaluated mentioned in the text.
+horizon_lengths_evaluated,,,not_reported,No specific horizon lengths evaluated mentioned in the text.
+inference_batch_size,,,not_reported,No inference batch size specified in the text or tables.
diff --git a/result/per_paper/2602.19068/accuracy_efficiency.csv b/result/per_paper/2602.19068/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..ffa78b1e4f729712bedf5ce759d7cd61f72f3759
--- /dev/null
+++ b/result/per_paper/2602.19068/accuracy_efficiency.csv
@@ -0,0 +1,42 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,Anomaly Transformer,TimeRadar,AUC-R,49.06,
+accuracy,DCdetector,TimeRadar,AUC-R,50.19,
+accuracy,D3R,TimeRadar,AUC-R,45.77,
+accuracy,ModernTCN,TimeRadar,AUC-R,63.04,
+accuracy,CATCH,TimeRadar,AUC-R,65.36,
+accuracy,GPT4TS,TimeRadar,AUC-R,58.85,
+accuracy,Zero-shot,TimeRadar,AUC-R,62.80,
+accuracy,Chronos-Bolt,TimeRadar,AUC-R,67.14,
+accuracy,Time-MoE,TimeRadar,AUC-R,61.04,
+accuracy,SEMPO,TimeRadar,AUC-R,73.46,
+accuracy,DADA,TimeRadar,AUC-R,75.15,
+accuracy,TimeRadar,TimeRadar,AUC-R,76.02,
+accuracy,Anomaly Transformer,TimeRadar,AUC-R,10.47,
+accuracy,DCdetector,TimeRadar,AUC-R,11.01,
+accuracy,D3R,TimeRadar,AUC-R,09.45,
+accuracy,ModernTCN,TimeRadar,AUC-R,15.20,
+accuracy,CATCH,TimeRadar,AUC-R,16.34,
+accuracy,GPT4TS,TimeRadar,AUC-R,13.83,
+accuracy,Zero-shot,TimeRadar,AUC-R,07.37,
+accuracy,Chronos-Bolt,TimeRadar,AUC-R,18.72,
+accuracy,Time-MoE,TimeRadar,AUC-R,16.79,
+accuracy,SEMPO,TimeRadar,AUC-R,22.04,
+accuracy,DADA,TimeRadar,AUC-R,26.42,
+accuracy,TimeRadar,TimeRadar,AUC-R,24.38,
+accuracy,Anomaly Transformer,TimeRadar,AUC-R,66.49,
+accuracy,DCdetector,TimeRadar,AUC-R,70.56,
+accuracy,D3R,TimeRadar,AUC-R,77.02,
+accuracy,ModernTCN,TimeRadar,AUC-R,77.17,
+accuracy,CATCH,TimeRadar,AUC-R,70.97,
+accuracy,GPT4TS,TimeRadar,AUC-R,77.23,
+accuracy,Zero-shot,TimeRadar,AUC-R,72.52,
+accuracy,Chronos-Bolt,TimeRadar,AUC-R,75.78,
+accuracy,Time-MoE,TimeRadar,AUC-R,66.05,
+accuracy,SEMPO,TimeRadar,AUC-R,74.45,
+accuracy,DADA,TimeRadar,AUC-R,78.27,
+accuracy,TimeRadar,TimeRadar,AUC-R,79.44,
+accuracy,A.1 Frequency domain + CDL,TimeRadar,"Aff-F1, AUC-R, AUC-P",1,
+accuracy,A.2 Time domain + CDL,TimeRadar,"Aff-F1, AUC-R, AUC-P",2,
+memory,Model Size,TimeRadar,Model Size,1.6,
+memory,Dataset Scale,TimeRadar,Model Size,408,
+memory,Context Length,TimeRadar,Model Size,100,
diff --git a/result/per_paper/2602.19068/accuracy_efficiency_traced.csv b/result/per_paper/2602.19068/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..0e7a87787b43e30eb1ffd8e6c553d71c9bba23e3
--- /dev/null
+++ b/result/per_paper/2602.19068/accuracy_efficiency_traced.csv
@@ -0,0 +1,42 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,Anomaly Transformer,TimeRadar,AUC-R,49.06,,1,2,2
+accuracy,DCdetector,TimeRadar,AUC-R,50.19,,1,3,2
+accuracy,D3R,TimeRadar,AUC-R,45.77,,1,4,2
+accuracy,ModernTCN,TimeRadar,AUC-R,63.04,,1,5,2
+accuracy,CATCH,TimeRadar,AUC-R,65.36,,1,6,2
+accuracy,GPT4TS,TimeRadar,AUC-R,58.85,,1,7,2
+accuracy,Zero-shot,TimeRadar,AUC-R,62.80,,1,8,2
+accuracy,Chronos-Bolt,TimeRadar,AUC-R,67.14,,1,9,2
+accuracy,Time-MoE,TimeRadar,AUC-R,61.04,,1,10,2
+accuracy,SEMPO,TimeRadar,AUC-R,73.46,,1,11,2
+accuracy,DADA,TimeRadar,AUC-R,75.15,,1,12,2
+accuracy,TimeRadar,TimeRadar,AUC-R,76.02,,1,13,2
+accuracy,Anomaly Transformer,TimeRadar,AUC-R,10.47,,1,15,2
+accuracy,DCdetector,TimeRadar,AUC-R,11.01,,1,16,2
+accuracy,D3R,TimeRadar,AUC-R,09.45,,1,17,2
+accuracy,ModernTCN,TimeRadar,AUC-R,15.20,,1,18,2
+accuracy,CATCH,TimeRadar,AUC-R,16.34,,1,19,2
+accuracy,GPT4TS,TimeRadar,AUC-R,13.83,,1,20,2
+accuracy,Zero-shot,TimeRadar,AUC-R,07.37,,1,21,2
+accuracy,Chronos-Bolt,TimeRadar,AUC-R,18.72,,1,22,2
+accuracy,Time-MoE,TimeRadar,AUC-R,16.79,,1,23,2
+accuracy,SEMPO,TimeRadar,AUC-R,22.04,,1,24,2
+accuracy,DADA,TimeRadar,AUC-R,26.42,,1,25,2
+accuracy,TimeRadar,TimeRadar,AUC-R,24.38,,1,26,2
+accuracy,Anomaly Transformer,TimeRadar,AUC-R,66.49,,1,28,2
+accuracy,DCdetector,TimeRadar,AUC-R,70.56,,1,29,2
+accuracy,D3R,TimeRadar,AUC-R,77.02,,1,30,2
+accuracy,ModernTCN,TimeRadar,AUC-R,77.17,,1,31,2
+accuracy,CATCH,TimeRadar,AUC-R,70.97,,1,32,2
+accuracy,GPT4TS,TimeRadar,AUC-R,77.23,,1,33,2
+accuracy,Zero-shot,TimeRadar,AUC-R,72.52,,1,34,2
+accuracy,Chronos-Bolt,TimeRadar,AUC-R,75.78,,1,35,2
+accuracy,Time-MoE,TimeRadar,AUC-R,66.05,,1,36,2
+accuracy,SEMPO,TimeRadar,AUC-R,74.45,,1,37,2
+accuracy,DADA,TimeRadar,AUC-R,78.27,,1,38,2
+accuracy,TimeRadar,TimeRadar,AUC-R,79.44,,1,39,2
+accuracy,A.1 Frequency domain + CDL,TimeRadar,"Aff-F1, AUC-R, AUC-P",1,,3,3,0
+accuracy,A.2 Time domain + CDL,TimeRadar,"Aff-F1, AUC-R, AUC-P",2,,3,4,0
+memory,Model Size,TimeRadar,Model Size,1.6,,5,2,1
+memory,Dataset Scale,TimeRadar,Model Size,408,,5,5,1
+memory,Context Length,TimeRadar,Model Size,100,,5,6,1
diff --git a/result/per_paper/2602.19068/components_architecture.csv b/result/per_paper/2602.19068/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..4259dec636b6375b63923750437a1bf323297c8a
--- /dev/null
+++ b/result/per_paper/2602.19068/components_architecture.csv
@@ -0,0 +1,3 @@
+component,what_it_is,provenance,citation,evidence
+Fractionally modulated Time-Frequency Reconstruction (FTFRecon),"A component that leverages a learnable fractional order to rotate time series into an optimal fractional time–frequency domain for adaptive data reconstruction, enabling differentiation of normal and abnormal patterns.",proposed_here,,"A novel component, namely Fractionally modulated Time-Frequency Reconstruction (FTFRecon), is proposed in TimeRadar to leverage a learnable fractional order to rotate the time series to the most pronounced angle between a continuous time and frequency domain for accurate data reconstruction."
+Contextual Deviation Learning (CDL),"A component that models patch-wise deviation of each input relative to its contextual time series data, promoting compactness among normal samples and increasing separation from anomalies.",proposed_here,,"To capture local abnormalities that global data reconstruction fails to characterize adequately, we further introduce a Contextual Deviation Learning (CDL) component in the rotatable domain, which models the patch-wise deviation of each input with respect to its contextual time series data, promoting compactness among normal samples while enlarging their separation from anomalies with a clear margin."
diff --git a/result/per_paper/2602.19068/computational.csv b/result/per_paper/2602.19068/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..6a354168d961878f9606d293581de9e6e7ae0dc1
--- /dev/null
+++ b/result/per_paper/2602.19068/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,Pro6000-96G GPU,GPU,stated,The entire pre-training process takes 3 hours on 2 Pro6000-96G GPUs...
+num_devices,2,units,stated,The entire pre-training process takes 3 hours on 2 Pro6000-96G GPUs...
+training_cost,3,hours,stated,The entire pre-training process takes 3 hours on 2 Pro6000-96G GPUs...
+training_batch_size,2048,samples,stated,"with a batch size of 2,048"
+training_steps_or_epochs,,not_reported,not_reported,not_reported
+precision,BF32,precision,stated,with BF32 precision
+inference_latency,29,seconds,stated,Inference Time (Seconds): ~29
+inference_throughput,,not_reported,not_reported,not_reported
+peak_memory,,not_reported,not_reported,not_reported
+flops_or_macs,,not_reported,not_reported,not_reported
+num_inference_samples,,not_reported,not_reported,not_reported
+params,1.6,M,stated,Model Size: 1.6M
+context_lengths_evaluated,,not_reported,not_reported,not_reported
+horizon_lengths_evaluated,,not_reported,not_reported,not_reported
+inference_batch_size,128,samples,stated,All experiments are conducted on a single Pro6000 GPU with a batch size of 128.
diff --git a/result/per_paper/2603.04791/accuracy_efficiency.csv b/result/per_paper/2603.04791/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2603.04791/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2603.04791/accuracy_efficiency_traced.csv b/result/per_paper/2603.04791/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2603.04791/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2603.04791/components_architecture.csv b/result/per_paper/2603.04791/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..2d0bbfc3197a60f65189cab3c8c56c1c5de28bf5
--- /dev/null
+++ b/result/per_paper/2603.04791/components_architecture.csv
@@ -0,0 +1,6 @@
+component,what_it_is,provenance,citation,evidence
+TimeMoE Block,Sparse Mixture-of-Experts (MoE) blocks integrated into Timer-S1 to enable efficient parameter activation (0.75B per token).,proposed_here,,Timer-S1 integrates sparse TimeMoE blocks and generic TimeSTP blocks for Serial-Token Prediction (STP).
+TimeSTP Block,"A serialized Transformer block for Serial-Token Prediction (STP), enabling progressive serial computations for multi-horizon forecasts.",proposed_here,,"Serial forecasting is implemented through a generic TimeSTP block for Serial-Token Prediction (STP), a serialized version of the Transformer block."
+Transformer Backbone,"The core architecture of Timer-S1, adapted for time series forecasting with autoregressive and multi-dimensional modeling capabilities.",reused_cited,"Su et al., 2023 [35]",Timer-XL addressed structural variability through a generic self-attention mechanism capable of modeling multi-dimensional time series in an autoregressive approach.
+Forecasting Head,"The output layer responsible for generating time series predictions, trained with a weighted STP objective during pre-training.",proposed_here,,"The model is densely supervised, where a training sample becomes multiple forecasting tasks with variable input and output lengths."
+Normalization and Embedding Layer,"Input processing module for time series data, including normalization and embedding to handle distributional heterogeneity.",proposed_here,,Section 3.1: Normalization and Embedding.
diff --git a/result/per_paper/2603.04791/computational.csv b/result/per_paper/2603.04791/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..aa200ef3c769cb8ad53e016a30dcf78fa5a4ff06
--- /dev/null
+++ b/result/per_paper/2603.04791/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No specific hardware type (e.g., GPU/TPU) is mentioned."
+num_devices,,,not_reported,No information on the number of devices used for training or inference.
+training_cost,,,not_reported,The text mentions 'lower the overall training cost' but does not quantify it.
+training_batch_size,,,not_reported,No explicit value for training batch size is provided.
+training_steps_or_epochs,,,not_reported,No mention of training steps or epochs.
+precision,,,not_reported,"No precision (e.g., FP16, FP32) is specified."
+inference_latency,,,not_reported,No latency metrics for inference are reported.
+inference_throughput,,,not_reported,No throughput metrics for inference are reported.
+peak_memory,,,not_reported,No peak memory usage is mentioned.
+flops_or_macs,,,not_reported,No FLOPs or MACs are reported.
+num_inference_samples,,,not_reported,No number of inference samples is specified.
+params,8300000000,parameters,stated,"The text states: 'Timer-S1, a strong Mixture-of-Experts (MoE) time series foundation model with 8.3B total parameters.'"
+context_lengths_evaluated,11500,tokens,stated,"The text states: 'a context length of 11.5K' (interpreted as 11,500 tokens)."
+horizon_lengths_evaluated,272,tokens,stated,The text states: 'prediction length $(H + 1)P = 272$' where $H = 16$ and $P = 16$.
+inference_batch_size,,,not_reported,No inference batch size is mentioned.
diff --git a/result/per_paper/2605.20119/accuracy_efficiency.csv b/result/per_paper/2605.20119/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..a6f1bf2aa149721fe94280102c3b15e63eaecf53
--- /dev/null
+++ b/result/per_paper/2605.20119/accuracy_efficiency.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
diff --git a/result/per_paper/2605.20119/accuracy_efficiency_traced.csv b/result/per_paper/2605.20119/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..be75ae07ef49c39cd9631c199954df8280a5a64c
--- /dev/null
+++ b/result/per_paper/2605.20119/accuracy_efficiency_traced.csv
@@ -0,0 +1 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
diff --git a/result/per_paper/2605.20119/components_architecture.csv b/result/per_paper/2605.20119/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..9a8a8ebd14344e6b3a38efe89c44e54117a8df98
--- /dev/null
+++ b/result/per_paper/2605.20119/components_architecture.csv
@@ -0,0 +1,7 @@
+component,what_it_is,provenance,citation,evidence
+Contiguous Patch Masking (CPM),"A parallel decoding scheme that replaces autoregressive decoding, enabling single-pass forecasting by masking contiguous patches during training and inference.",reused_cited,"Auer et al., 2025","Toto 2.0 replaces Toto 1.0's autoregressive decoding with contiguous patch masking, an elegant single-pass parallel scheme adapted from Auer et al. (2025)."
+Quantile Output Head,"A regression head that outputs nine quantile levels to improve forecast stability at scale, replacing the Student-T mixture used in Toto 1.0.",proposed_here,,A quantile output head replaces the Student-T mixture of Toto 1.0 to improve stability at scale.
+Robust Causal Scaler,An input preprocessing module using arcsinh normalization to stabilize training and improve generalization.,proposed_here,,A robust causal scaler (arcsinh normalization) on the input side is part of the Toto 2.0 architecture.
+Variate-Time Transformer Decoder,"A decoder-only transformer with alternating time-axis (causal) and variate-axis (full) attention layers, retained from Toto 1.0.",reused_cited,"Cohen et al., 2024","The Toto 2.0 backbone is largely retained from Toto 1.0 (Cohen et al., 2024): a decoder-only patched transformer whose attention layers alternate between time-axis (causal) and variate-axis (full) views of the input."
+Input Residual MLP,A multi-layer perceptron applied after patch embedding to refine input representations before transformer processing.,proposed_here,,The forward pass includes an Input Residual MLP after patch embedding and masking.
+Output Residual MLP,A multi-layer perceptron applied after the transformer decoder to refine output representations before quantile prediction.,proposed_here,,The forward pass includes an Output Residual MLP before the quantile output head.
diff --git a/result/per_paper/2605.20119/computational.csv b/result/per_paper/2605.20119/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..28f33d76f67c8b08129286ac48c8798e8d44fca1
--- /dev/null
+++ b/result/per_paper/2605.20119/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,,,not_reported,"No specific hardware type (e.g., GPU, TPU) is mentioned for Toto 2.0."
+num_devices,,,not_reported,"The number of devices (e.g., GPUs) used for training or inference is not specified."
+training_cost,,,not_reported,No explicit mention of monetary or resource cost for training Toto 2.0.
+training_batch_size,64,samples,stated,The text states: 'All five sizes train on [...] at a global batch size of 64.'
+training_steps_or_epochs,,,not_reported,"Training steps for Toto 2.0 are not explicitly stated; the proxy model is trained for 30,000 steps, but this is not directly tied to Toto 2.0."
+precision,,,not_reported,"Low precision is mentioned in the context of optimization, but specific precision (e.g., 16-bit, 32-bit) is not reported."
+inference_latency,,,not_reported,No latency metrics for inference are provided.
+inference_throughput,,,not_reported,No throughput metrics for inference are provided.
+peak_memory,,,not_reported,Peak memory usage is not mentioned.
+flops_or_macs,,,not_reported,No FLOPs or MACs are reported for Toto 2.0.
+num_inference_samples,,,not_reported,The number of inference samples is not specified.
+params,,,not_reported,"Model sizes (e.g., 4m, 22m, 313m, 1B, 2.5B) are listed, but the text does not explicitly state the parameter count for Toto 2.0 itself."
+context_lengths_evaluated,"[2048, 4096, 8192]",timesteps,stated,"The text states: 'evaluated [...] at horizons of 2,048, 4,096, and 8,192 timesteps.'"
+horizon_lengths_evaluated,"[2048, 4096, 8192]",timesteps,stated,"The text states: 'evaluated [...] at horizons of 2,048, 4,096, and 8,192 timesteps.'"
+inference_batch_size,,,not_reported,"The training batch size is mentioned (64), but inference batch size is not specified."
diff --git a/result/per_paper/2606.10798/accuracy_efficiency.csv b/result/per_paper/2606.10798/accuracy_efficiency.csv
new file mode 100644
index 0000000000000000000000000000000000000000..0dd8ec0b86c4a821bb0748b45c370c4fd5ae2479
--- /dev/null
+++ b/result/per_paper/2606.10798/accuracy_efficiency.csv
@@ -0,0 +1,7 @@
+metric_kind,entity,model_variant,metric,value,uncertainty
+accuracy,$YINGLONG_{6m}$,CITRAS-FM,Skill Score,6,
+accuracy,TinyTimeMixer $_{r2}$,CITRAS-FM,Skill Score,2,
+accuracy,Chronos-2,CITRAS-FM,Skill Score,-2,
+accuracy,TimesFM-2.5,CITRAS-FM,Skill Score,-2.5,
+accuracy,Toto-1.0,CITRAS-FM,Skill Score,-1.0,
+accuracy,Moirai-2.0,CITRAS-FM,Skill Score,-2.0,
diff --git a/result/per_paper/2606.10798/accuracy_efficiency_traced.csv b/result/per_paper/2606.10798/accuracy_efficiency_traced.csv
new file mode 100644
index 0000000000000000000000000000000000000000..55b9722a3e0519f4060beb34117e2f92ca297d03
--- /dev/null
+++ b/result/per_paper/2606.10798/accuracy_efficiency_traced.csv
@@ -0,0 +1,7 @@
+metric_kind,entity,model_variant,metric,value,uncertainty,table,row,col
+accuracy,$YINGLONG_{6m}$,CITRAS-FM,Skill Score,6,,0,6,0
+accuracy,TinyTimeMixer $_{r2}$,CITRAS-FM,Skill Score,2,,0,7,0
+accuracy,Chronos-2,CITRAS-FM,Skill Score,-2,,0,10,0
+accuracy,TimesFM-2.5,CITRAS-FM,Skill Score,-2.5,,0,12,0
+accuracy,Toto-1.0,CITRAS-FM,Skill Score,-1.0,,0,13,0
+accuracy,Moirai-2.0,CITRAS-FM,Skill Score,-2.0,,0,17,0
diff --git a/result/per_paper/2606.10798/components_architecture.csv b/result/per_paper/2606.10798/components_architecture.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5329bd50c9452367ddb18273a6e8a8ae2d4d49ad
--- /dev/null
+++ b/result/per_paper/2606.10798/components_architecture.csv
@@ -0,0 +1,12 @@
+component,what_it_is,provenance,citation,evidence
+Shifted Attention,A novel attention mechanism in the cross-variate module that enables effective use of covariates throughout the forecast horizon.,proposed_here,,CITRAS-FM introduces a novel Shifted Attention layer in the cross-variate module of CITRAS to improve generalization in capturing covariate information.
+CovSynth,A method for synthesizing pseudo-covariates from decomposed components of target series to enable covariate-aware pretraining.,proposed_here,,"To address the scarcity of covariates in existing pretraining corpora, we propose CovSynth, a method that synthesizes pseudo-covariates that capture target fluctuations using decomposed components of the target series."
+Patch-based Decoder-only Transformer,"The foundational architecture of CITRAS-FM, derived from CITRAS [8], designed for time series forecasting with patch-based tokenization.",reused_cited,"Yamaguchi et al., 2023 [8]","CITRAS-FM builds on our prior work, CITRAS [8], which enables flexible handling of different variable types under supervised settings."
+Input Projection,A module that converts raw time series into token embeddings using patching and causal scaling.,reused_cited,"Yamaguchi et al., 2023 [8]","This module converts raw time series into token embeddings. To preserve the local semantics of time series, we first apply patching [20], where each variable is segmented into non-overlapping patches of length P."
+Cross-Time Attention,A module that captures temporal dependencies within a single time series.,reused_cited,"Yamaguchi et al., 2023 [8]","The architecture consists of four modules: input projection, cross-time attention, cross-variate attention, and output projection."
+Cross-Variate Attention,A module that models relationships between different variables (including covariates) using Shifted Attention.,reused_cited,"Yamaguchi et al., 2023 [8]",The cross-variate attention module incorporates Shifted Attention to effectively exploit known covariates.
+Output Projection,A module that maps transformer outputs to forecasted time series values.,reused_cited,"Yamaguchi et al., 2023 [8]","The architecture consists of four modules: input projection, cross-time attention, cross-variate attention, and output projection."
+Quantile Regression Head,A module for probabilistic forecasting that outputs quantile-based predictions.,proposed_here,,We incorporate a quantile regression head for probabilistic forecasting.
+Causal Scaling,A technique to mitigate non-stationarity while preventing information leakage from future patches.,reused_cited,TimesFM-2.5 [3],We apply patch-wise causal scaling [3] to mitigate non-stationarity while preventing information leakage from future patches.
+Pre-Layer Normalization,A normalization technique applied before each layer to improve training stability.,reused_cited,Transformer [18],We incorporate recent techniques such as pre-layer normalization [18] for improved stability.
+SwiGLU Activation,A non-linear activation function used in the feed-forward network for improved model expressiveness.,reused_cited,SwiGLU [19],We incorporate recent techniques such as SwiGLU activation [19] for improved stability.
diff --git a/result/per_paper/2606.10798/computational.csv b/result/per_paper/2606.10798/computational.csv
new file mode 100644
index 0000000000000000000000000000000000000000..064e823612f8358613083a359e3ce3a748d5bc83
--- /dev/null
+++ b/result/per_paper/2606.10798/computational.csv
@@ -0,0 +1,16 @@
+field,value,unit,source,evidence
+hardware_type,NVIDIA V100 (32 GB),,stated,A single NVIDIA V100 (32 GB) GPU is used for pretraining.
+num_devices,1,,stated,A single NVIDIA V100 (32 GB) GPU is used for pretraining.
+training_cost,,,not_reported,
+training_batch_size,256,,stated,pretraining ... batch size of 256
+training_steps_or_epochs,500000,,stated,"train for 500,000 steps"
+precision,,,not_reported,
+inference_latency,0.05,sec,stated,Inference Time (sec) ↓: 0.05
+inference_throughput,,,not_reported,
+peak_memory,,,not_reported,
+flops_or_macs,,,not_reported,
+num_inference_samples,,,not_reported,
+params,7.2,M,stated,CITRAS-FM | 7.2 | ... (Params. (M))
+context_lengths_evaluated,1032,,stated,"fix the historical input length to 1,032 for all models"
+horizon_lengths_evaluated,24,,stated,forecasting horizon is 24 steps
+inference_batch_size,,,not_reported,