Update README.md

updated data card

README.md

@@ -1,6 +1,7 @@
 ---
 pretty_name: mTSBench
 license: cc-by-nc-4.0
+arxiv: "2506.21550"
 size_categories:
 - 10M<n<100M
 tags:
@@ -165,22 +166,90 @@ configs:
     path: swan/*_test.csv
 ---
 
-# mTSBench
+# mTSBench: Benchmarking Multivariate Time Series Anomaly Detection and Model Selection at Scale
 
-mTSBench is a collection of 344 multivariate time series from 19 datasets commonly used in anomaly detection research. Each folder corresponds to one dataset and contains `*_train.csv`, `*_test.csv`, and `*_val.csv` files. See [`data_summary.csv`](https://github.com/PLAN-Lab/mTSBench/blob/main/Datasets/data_summary.csv) for per-file statistics.
+**Xiaona Zhou, Constantin Brif, Ismini Lourentzou**
 
-## How to download
+### 🎉 Accepted to TMLR 2026
 
-This repository uses Git LFS for the CSV files.
+[📄 Paper](https://openreview.net/forum?id=8LfB8HD1WU) · [📝 arXiv](https://arxiv.org/abs/2506.21550) · [🌐 Project Page](https://plan-lab.github.io/projects/mtsbench/) · [💻 Code](https://github.com/PLAN-Lab/mTSBench)
+
+---
+
+## Overview
+
+**mTSBench** is the largest and most diverse benchmark for multivariate time series anomaly detection (MTS-AD) and model selection. It contains **344 labeled multivariate time series from 19 datasets across 12 application domains**, and evaluates **24 anomaly detectors** and **3 unsupervised model selection methods** under a unified evaluation suite.
+
+> **No single anomaly detector performs consistently well across datasets, and current unsupervised model selection methods remain far from optimal.**
+
+<p align="center">
+  <img src="https://plan-lab.github.io/projects/mtsbench/static/images/mTSBench.png" width="100%">
+</p>
+
+---
+
+## Key Idea
+
+Multivariate time series anomaly detection is challenging because of high-dimensional dependencies, cross-correlations between time-dependent variables, and the scarcity of labeled anomalies. Because no single detector dominates across datasets, choosing the right detector for a new time series is a problem in its own right.
+
+**mTSBench** is the first benchmark to integrate unsupervised model selection methods and evaluate them alongside anomaly detectors under consistent settings, with a unified evaluation suite of point-based and ranking-based metrics.
+
+---
+
+## Highlights
+
+* **Largest MTS-AD benchmark:** 344 labeled multivariate time series from 19 datasets across 12 application domains, including healthcare, cybersecurity, industrial monitoring, spacecraft telemetry, and IT infrastructure.
+
+* **24 anomaly detectors:** Spanning reconstruction-, prediction-, statistics-, and LLM-based approaches, including the only two publicly available LLM-based methods for multivariate time series.
+
+* **3 unsupervised model selection methods:** MetaOD, FMMS, and Orthus.
+
+* **Unified evaluation suite:** Point-based and ranking-based metrics for anomaly detection and model selection, including VUS-PR, VUS-ROC, AUC-PR, AUC-ROC, AUC-PTRT, and Precision@k, Recall@k, and NDCG@k.
+
+* **Key finding:** Even the strongest selectors remain far below the near-optimal baseline, and a single detector, PCA, surpasses all three selectors on 9 of 13 metrics.
+
+---
+
+## Dataset Overview
+
+| Dataset | Config | Domain | #TS | #Dim | Length | #AnomPts | #AnomSeqs | License |
+|---|---|---|---:|---:|---:|---:|---:|---|
+| CIC-IDS-2017 | `cicids` | Cybersecurity | 5 | 73 | > 100K | 0–8656 | 0–2546 | Citation Required |
+| CalIt2 | `CalIt2` | Smart Building | 1 | 3 | > 5K | 0 | 21 | CC BY 4.0 |
+| CreditCard | `creditcard` | Finance / Fraud Detection | 1 | 30 | > 100K | 219 | 10 | Citation Required |
+| Daphnet | `Daphnet` | Healthcare | 26 | 10 | > 50K | 0 | 1–16 | CC BY 4.0 |
+| Exathlon | `Exathlon` | IT Infrastructure | 30 | 21 | > 50K | 0–4 | 0–6 | Apache 2.0 |
+| GECCO | `GECCO` | Industrial Process | 1 | 10 | > 50K | 0 | 37 | Citation Required |
+| GHL | `GHL` | Industrial Process | 14 | 17 | > 100K | 0 | 1–4 | Contact Authors |
+| Genesis | `Genesis` | Industrial Process | 1 | 19 | > 5K | 0 | 2 | CC BY-NC-SA 4.0 |
+| GutenTAG | `GutenTAG` | Synthetic Benchmark | 30 | 21 | > 10K | 0 | 1–3 | MIT |
+| MITDB | `MITDB` | Healthcare | 47 | 3 | > 500K | 0 | 1–720 | ODC-By v1.0 |
+| MSL | `MSL` | Spacecraft Telemetry | 26 | 56 | > 5K | 0 | 1–3 | BSD 3-Clause |
+| Metro | `metro` | Transportation | 1 | 6 | > 10K | 20 | 5 | CC BY 4.0 |
+| OPPORTUNITY | `OPPORTUNITY` | Human Activity Recognition | 13 | 33 | > 25K | 0 | 1 | CC BY 4.0 |
+| Occupancy | `room-occupancy` | Smart Building | 2 | 6 | > 5K | 1–3 | 9–13 | CC BY 4.0 |
+| PSM | `PSM` | IT Infrastructure | 1 | 27 | > 50K | 0 | 39 | CC BY 4.0 |
+| SMAP | `SMAP` | Spacecraft Telemetry | 48 | 26 | > 5K | 0 | 1–3 | BSD 3-Clause |
+| SMD | `SMD` | IT Infrastructure | 18 | 39 | > 10K | 0 | 4–24 | MIT |
+| SVDB | `SVDB` | Healthcare | 78 | 3 | > 100K | 0 | 2–678 | ODC-By v1.0 |
+| SWAN-SF | `swan` | Astrophysics | 1 | 39 | > 50K | 5233 | 1382 | MIT |
+
+#AnomPts and #AnomSeqs are the numbers of anomalous points and anomalous sequences per time series. Licenses are those of the original datasets.
+
+Each folder corresponds to one dataset and contains `*_train.csv`, `*_val.csv`, and `*_test.csv` files. Each CSV contains a `timestamp` column, dataset-specific feature columns, and a binary `is_anomaly` label. The `*_val.csv` files are held-out validation time series used to train the model selection methods. See [`data_summary.csv`](https://github.com/PLAN-Lab/mTSBench/blob/main/Datasets/data_summary.csv) for per-file statistics.
+
+---
+
+## Usage
+
+This repository uses Git LFS for the CSV files:
 
 ```bash
 git lfs install
 git clone https://huggingface.co/datasets/PLAN-Lab/mTSBench
 ```
 
-## Load with Hugging Face
-
-Select one of the 19 configurations so files with different schemas are not combined.
+Or load one of the 19 configurations with 🤗 Datasets (one configuration per dataset, so files with different schemas are not combined):
 
 ```python
 from datasets import load_dataset
@@ -191,34 +260,21 @@ df_train = calit2["train"].to_pandas()
 df_test = calit2["test"].to_pandas()
 ```
 
-Each CSV contains a `timestamp` column, dataset-specific feature columns, and a binary `is_anomaly` label.
-
-## Dataset overview
-
-| Dataset | Domain | #TS | #Dims | Length | #AnomPts | #AnomSeqs |
-|---|---|---:|---:|---:|---:|---:|
-| CalIt2 | Smart Building | 1 | 3 | >5K | 0 | 21 |
-| CreditCard | Finance / Fraud Detection | 1 | 30 | >100K | 219 | 10 |
-| Daphnet | Healthcare | 26 | 10 | >50K | 0 | 1–16 |
-| Exathlon | Cloud Computing | 30 | 21 | >50K | 0–4 | 0–6 |
-| GECCO | Water Quality Monitoring | 1 | 10 | >50K | 0 | 37 |
-| GHL | Industrial Process | 14 | 17 | >100K | 0 | 1–4 |
-| Genesis | Industrial Automation | 1 | 19 | >5K | 0 | 2 |
-| GutenTAG | Synthetic Benchmark | 30 | 21 | >10K | 0 | 1–3 |
-| MITDB | Healthcare | 47 | 3 | >500K | 0 | 1–720 |
-| MSL | Spacecraft Telemetry | 26 | 56 | >5K | 0 | 1–3 |
-| OPPORTUNITY | Human Activity Recognition | 13 | 33 | >25K | 0 | 1 |
-| Occupancy | Smart Building | 2 | 6 | >5K | 1–3 | 9–13 |
-| PSM | IT Infrastructure | 1 | 27 | >50K | 0 | 39 |
-| SMAP | Spacecraft Telemetry | 48 | 26 | >5K | 0 | 1–3 |
-| SMD | IT Infrastructure | 18 | 39 | >10K | 0 | 4–24 |
-| SVDB | Healthcare | 78 | 3 | >100K | 0 | 2–678 |
-| CIC-IDS-2017 | Cybersecurity | 5 | 73 | >100K | 0–8656 | 0–2546 |
-| Metro | Transportation | 1 | 6 | >10K | 20 | 5 |
-| SWAN-SF | Industrial Process | 1 | 39 | >50K | 5233 | 1382 |
+---
+
+## Resources
+
+* **Paper:** [mTSBench: Benchmarking Multivariate Time Series Anomaly Detection and Model Selection at Scale](https://openreview.net/forum?id=8LfB8HD1WU) (TMLR 2026)
+* **arXiv:** [arxiv.org/abs/2506.21550](https://arxiv.org/abs/2506.21550)
+* **Project Page:** [plan-lab.github.io/projects/mtsbench](https://plan-lab.github.io/projects/mtsbench/)
+* **Code:** [github.com/PLAN-Lab/mTSBench](https://github.com/PLAN-Lab/mTSBench)
+
+---
 
 ## Citation
 
+If you find mTSBench useful in your research, please consider citing our work:
+
 ```bibtex
 @article{zhou2026mtsbench,
   title={mTSBench: Benchmarking Multivariate Time Series Anomaly Detection and Model Selection at Scale},
@@ -227,10 +283,3 @@ Each CSV contains a `timestamp` column, dataset-specific feature columns, and a
   year={2026}
 }
 ```
-
-## Links
-
-- [Paper](https://arxiv.org/abs/2506.21550)
-- [Code](https://github.com/PLAN-Lab/mTSBench)
-- [Project page](https://plan-lab.github.io/projects/mtsbench/)
-