Dataset Viewer
Auto-converted to Parquet Duplicate
all_domains
listlengths
0
13
all_methods
listlengths
0
15
all_novelty_signals
listlengths
3
6
all_query_ids
listlengths
0
8
archived
bool
2 classes
candidate_status
stringclasses
1 value
candidate_rule_version
stringclasses
1 value
candidate_eligible
bool
1 class
candidate_reason
stringclasses
1 value
created_at
stringdate
2011-10-23 16:15:22
2026-09-29 13:40:55
description
stringlengths
26
598
domains
listlengths
0
10
evidence_signals
listlengths
0
5
evidence_tier
stringclasses
3 values
evidence_version
stringclasses
1 value
first_observed_at
stringclasses
36 values
fork
bool
1 class
forks
int64
0
3.39k
github_id
int64
2.63M
1.4B
homepage
stringlengths
17
187
⌀
language
stringclasses
19 values
license
stringclasses
9 values
methods
listlengths
0
9
name
stringlengths
8
120
novelty_signals
listlengths
3
6
observation_count
int64
1
12
observed_at
stringclasses
32 values
paper_ids
listlengths
0
0
pushed_at
stringdate
2011-10-23 16:20:23
2026-09-29 13:47:16
query_ids
listlengths
0
4
readme_blob_sha
stringclasses
157 values
readme_checked_at
stringdate
2026-09-25 22:08:11
2026-09-29 18:47:14
⌀
readme_evidence_version
stringclasses
2 values
readme_etag
stringclasses
157 values
readme_sections
listlengths
0
9
⌀
readme_signals
listlengths
0
6
⌀
readme_status
stringclasses
2 values
readme_observed_at
stringdate
2026-09-25 22:08:11
2026-09-29 18:47:14
⌀
readme_repository_name_at_fetch
stringclasses
161 values
selection_reason
stringclasses
3 values
selection_signals
listlengths
2
8
selection_status
stringclasses
1 value
selection_version
stringclasses
1 value
stars
int64
0
35.7k
topics
listlengths
0
20
updated_at
stringdate
2013-12-14 17:24:37
2026-09-29 16:15:29
url
stringlengths
27
139
extra_json
stringclasses
21 values
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[ "neural-network-pruning", "pruning", "structured-pruning" ]
[ "description", "query-match", "repository-metadata" ]
[ "efficiency.pruning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2011-10-23T16:15:22Z
Algorithm MONSA for All Closed Sets Finding. MONSA is an exact depth-first search algorithm extracting only frequent closed sets using several new very effective pruning techniques to be free from repetitive and empty patterns.
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
0
2,631,335
null
C++
null
[ "neural-network-pruning", "pruning", "structured-pruning" ]
mikksoone/monsa
[ "description", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2011-10-23T16:20:23Z
[ "efficiency.pruning" ]
null
2026-09-29T18:47:14Z
gh-ml-readme-evidence-v2
null
[]
[]
missing
2026-09-29T18:47:14Z
mikksoone/monsa
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-cue" ]
include
ml-contribution-v5
1
[]
2013-12-14T17:24:37Z
https://github.com/mikksoone/monsa
null
[ "general-ml", "representation-learning" ]
[ "fine-tuning", "transfer-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.transfer-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2015-06-14T12:40:31Z
Code for paper "A Machine Learning Approach for Instance Matching Based on Similarity Metrics" ISWC2012.
[ "general-ml", "representation-learning" ]
[ "machine-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
0
37,411,816
null
Java
null
[ "fine-tuning", "transfer-learning" ]
FicusRong/Instance-Matching-by-Transfer-Learning
[ "description", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2015-06-15T07:12:24Z
[ "general.transfer-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-context-only", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
5
[]
2018-05-26T17:43:38Z
https://github.com/FicusRong/Instance-Matching-by-Transfer-Learning
null
[ "classical-ml", "tabular-ml" ]
[ "ensemble-learning", "random-forest" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.random-forest" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2015-10-21T02:26:09Z
code for paper "Feature-Budgeted Random Forest" ICML 2015
[ "classical-ml", "tabular-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
5
44,647,664
null
C++
MIT
[ "ensemble-learning", "random-forest" ]
fnan/FeatureBudgetedRandomForest
[ "description", "license-metadata", "query-match", "repository-metadata" ]
4
2026-09-28T18:50:00Z
[]
2017-05-10T13:37:36Z
[ "general.random-forest" ]
1b62dc8ab6b7f250ecd2b20fca7cee1ad83fb256
2026-09-26T11:23:39Z
gh-ml-readme-evidence-v1
"1b62dc8ab6b7f250ecd2b20fca7cee1ad83fb256"
[ "installation", "other", "usage" ]
[ "course-cue", "ml-method-context", "paper-reference", "survey-cue" ]
ok
2026-09-26T11:23:39Z
fnan/FeatureBudgetedRandomForest
official-paper-method-implementation
[ "course-cue", "ml-method-context", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue", "paper-reference", "survey-cue" ]
include
ml-contribution-v5
11
[]
2025-01-03T15:50:41Z
https://github.com/fnan/FeatureBudgetedRandomForest
null
[ "computational-neuroscience", "control", "machine-learning", "robotics", "robotics-and-control" ]
[ "control", "neural-network", "neuromorphic-computing", "spiking-neural-network" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "robotics.robot-control", "specialized.spiking-neural-network" ]
true
candidate
ml-candidate-v3
true
selected-by-current-rule
2016-01-15T14:16:57Z
Diverse, Noisy and Parallel: a New Spiking Neural Network Approach for Humanoid Robot Control
[ "computational-neuroscience", "machine-learning", "robotics-and-control" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
5
49,722,881
http://ieeexplore.ieee.org/document/7727325/
Jupyter Notebook
null
[ "neural-network", "neuromorphic-computing", "spiking-neural-network" ]
ricardodeazambuja/IJCNN2016
[ "description", "github-topics", "query-match", "repository-metadata" ]
3
2026-09-26T14:50:46Z
[]
2021-07-14T09:07:50Z
[ "specialized.spiking-neural-network" ]
30cf7d12c3b0a6869d26ee10cd474a8c21bb0bdd
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
"30cf7d12c3b0a6869d26ee10cd474a8c21bb0bdd"
[ "abstract", "citation", "method", "other" ]
[ "ml-method-context" ]
ok
2026-09-26T15:53:33Z
ricardodeazambuja/IJCNN2016
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-context", "ml-method-cue" ]
include
ml-contribution-v5
17
[ "baxter-robot", "liquid-state-machines", "lsm", "robot", "snn", "spiking-neural-networks", "vrep-simulator" ]
2026-05-05T13:38:49Z
https://github.com/ricardodeazambuja/IJCNN2016
null
[ "classical-ml" ]
[ "kernel-methods", "support-vector-machine" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.support-vector-machine" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2016-01-15T17:45:31Z
k-support regularized Support Vector Machine (ksup-SVM) is a novel regularization method that extends the L1 regularized SVM to a mixed norm of both L1 and L2 norms. This enables the use of a correlated sparsity regularization with the power of the SVM framework.
[ "classical-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-28T18:50:00Z
false
1
49,734,905
null
Matlab
GPL-3.0
[ "kernel-methods", "support-vector-machine" ]
gkirtzou/ksup_svm
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-28T18:50:00Z
[]
2016-01-15T17:46:26Z
[ "general.support-vector-machine" ]
1bea0d6a666fdb0f5db3638e5f57fe9e8d0da40f
2026-09-28T20:05:39Z
gh-ml-readme-evidence-v2
"1bea0d6a666fdb0f5db3638e5f57fe9e8d0da40f"
[]
[ "ml-method-context" ]
ok
2026-09-28T20:05:39Z
gkirtzou/ksup_svm
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-context", "ml-method-cue" ]
include
ml-contribution-v5
1
[]
2017-09-16T12:34:24Z
https://github.com/gkirtzou/ksup_svm
null
[ "general-ml", "reinforcement-learning" ]
[ "deep-reinforcement-learning", "paper-implementation", "policy-learning", "reinforcement-learning" ]
[ "description", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
[ "general.arxiv", "rl.deep" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2016-05-09T12:59:18Z
A Tensorflow based implementation of "Asynchronous Methods for Deep Reinforcement Learning": https://arxiv.org/abs/1602.01783
[ "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
23
58,376,719
null
Python
Apache-2.0
[ "deep-reinforcement-learning", "policy-learning", "reinforcement-learning" ]
traai/async-deep-rl
[ "description", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
2
2026-09-26T10:28:13Z
[]
2016-10-28T11:29:21Z
[ "rl.deep" ]
2f4599ca9126ba8b0c1aeed9e61dd5c58f68eb5a
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
"2f4599ca9126ba8b0c1aeed9e61dd5c58f68eb5a"
[ "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-26T15:53:33Z
traai/async-deep-rl
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-method-context", "ml-method-cue", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
68
[]
2025-09-12T07:50:25Z
https://github.com/traai/async-deep-rl
null
[ "computer-vision", "information-retrieval" ]
[ "image-retrieval", "metric-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "vision.image-retrieval" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2016-05-20T21:40:55Z
Code for paper Sketch Me That Shoe
[ "computer-vision", "information-retrieval" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
26
59,328,028
null
Jupyter Notebook
null
[ "image-retrieval", "metric-learning" ]
seuliufeng/DeepSBIR
[ "description", "query-match", "repository-metadata" ]
2
2026-09-26T10:28:13Z
[]
2018-04-27T16:40:53Z
[ "vision.image-retrieval" ]
df29b346cf7aaa48bfea796efb36ddc3c40ec56e
2026-09-26T20:25:07Z
gh-ml-readme-evidence-v2
"df29b346cf7aaa48bfea796efb36ddc3c40ec56e"
[ "other" ]
[ "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-26T20:25:07Z
seuliufeng/DeepSBIR
readme-supported-paper-method-implementation
[ "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
64
[]
2026-03-31T03:09:51Z
https://github.com/seuliufeng/DeepSBIR
null
[ "audio", "computational-neuroscience", "machine-learning", "speech-and-audio" ]
[ "audio-classification", "neural-network", "neuromorphic-computing", "representation-learning", "spiking-neural-network" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "audio-audio-classification", "specialized.spiking-neural-network" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2016-06-17T09:07:33Z
This is the PyNN code used in the paper titled "Multilayer Spiking Neural Network for audio samples classification using SpiNNaker", which is already accepted for publication.
[ "computational-neuroscience", "machine-learning", "speech-and-audio" ]
[ "neural-network", "classifier" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
10
61,360,870
null
Python
GPL-3.0
[ "neural-network", "neuromorphic-computing", "spiking-neural-network" ]
jpdominguez/Multilayer-SNN-for-audio-samples-classification-using-SpiNNaker
[ "description", "license-metadata", "query-match", "repository-metadata" ]
3
2026-09-26T14:50:46Z
[]
2021-12-07T10:07:17Z
[ "specialized.spiking-neural-network" ]
c4278e768503576fb2f0cd78e54d16cf5351258d
2026-09-26T20:25:07Z
gh-ml-readme-evidence-v2
"c4278e768503576fb2f0cd78e54d16cf5351258d"
[ "abstract", "citation", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-26T20:25:07Z
jpdominguez/Multilayer-SNN-for-audio-samples-classification-using-SpiNNaker
readme-supported-paper-method-implementation
[ "method-contribution", "ml-context-only", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
32
[]
2026-03-11T19:49:03Z
https://github.com/jpdominguez/Multilayer-SNN-for-audio-samples-classification-using-SpiNNaker
null
[ "deep-learning", "generative-modeling", "health-and-biomedicine", "medical-imaging", "probabilistic-ml" ]
[ "bayesian-deep-learning", "image-analysis", "uncertainty-estimation" ]
[ "description", "github-topics", "paper-reference", "query-match", "repository-metadata" ]
[ "general.bayesian-deep-learning", "medical.medical-imaging" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2016-11-30T12:09:59Z
Code for the paper "Improving Variational Auto-Encoders using Householder Flow" (https://arxiv.org/abs/1611.09630)
[ "generative-modeling", "health-and-biomedicine", "medical-imaging" ]
[ "deep-learning", "generative-model", "representation-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T15:29:21Z
false
12
75,183,533
https://jmtomczak.github.io/deebmed.html
Python
null
[ "image-analysis" ]
jmtomczak/vae_householder_flow
[ "description", "github-topics", "paper-reference", "query-match", "repository-metadata" ]
5
2026-09-28T18:50:00Z
[]
2017-01-26T09:18:13Z
[ "medical.medical-imaging" ]
528d3fffa7692dbfc08a6fe8290dd5b40cb18ccf
2026-09-26T20:42:46Z
gh-ml-readme-evidence-v2
"528d3fffa7692dbfc08a6fe8290dd5b40cb18ccf"
[ "citation", "other", "results" ]
[ "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-26T20:42:46Z
jmtomczak/vae_householder_flow
readme-supported-paper-method-implementation
[ "ml-context-only", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
74
[ "deep-learning", "generative-model", "normalizing-flows", "representation-learning", "variational-autoencoders" ]
2025-12-09T13:18:20Z
https://github.com/jmtomczak/vae_householder_flow
null
[ "general-ml", "generative-modeling" ]
[ "paper-implementation" ]
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
[ "general.arxiv" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-02-12T12:20:04Z
Tensorflow implementation of Wasserstein GAN - arxiv: https://arxiv.org/abs/1701.07875
[ "general-ml", "generative-modeling" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
129
81,723,773
null
Python
MIT
[ "paper-implementation" ]
shekkizh/WassersteinGAN.tensorflow
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
3
2026-09-24T16:52:35Z
[]
2017-02-13T20:49:15Z
[ "general.arxiv" ]
0a905a6db044bf0afed3788acf136f2d6625965b
2026-09-26T21:03:54Z
gh-ml-readme-evidence-v2
"0a905a6db044bf0afed3788acf136f2d6625965b"
[ "other", "references" ]
[ "method-contribution", "ml-method-context", "model-training-artifact", "paper-code-relationship", "paper-reference" ]
ok
2026-09-26T21:03:54Z
shekkizh/WassersteinGAN.tensorflow
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-method-context", "model-training-artifact", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
412
[ "gan", "generative-adversarial-network", "tensorflow", "wasserstein" ]
2026-07-15T06:37:30Z
https://github.com/shekkizh/WassersteinGAN.tensorflow
null
[ "earth-observation", "earth-science", "environmental-science", "geospatial", "geospatial-science", "remote-sensing", "science-and-engineering" ]
[ "deep-learning", "foundation-model", "geospatial-learning", "land-cover-classification", "machine-learning", "remote-sensing", "semantic-segmentation" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "geo.land-cover", "geo.topic-remote-sensing", "science.remote-sensing" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-03-05T19:49:26Z
Data and code for the paper "Remote Sensing-Based Measurement of Living Environment Deprivation - Improving Classical Approaches with Machine Learning", by Dani Arribas-Bel, Jorge Patiño and Juanca Duque
[ "earth-observation", "earth-science", "environmental-science", "geospatial", "geospatial-science", "remote-sensing", "science-and-engineering" ]
[ "machine-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
9
83,997,489
null
Jupyter Notebook
null
[ "deep-learning", "foundation-model", "geospatial-learning", "land-cover-classification", "machine-learning", "remote-sensing", "semantic-segmentation" ]
darribas/satellite_led_liverpool
[ "description", "github-topics", "query-match", "repository-metadata" ]
3
2026-09-26T10:28:13Z
[]
2019-03-13T10:53:44Z
[ "geo.land-cover", "geo.topic-remote-sensing", "science.remote-sensing" ]
1852140aef0a9d7900bae77e88b08a5e0cce01e1
2026-09-26T21:03:54Z
gh-ml-readme-evidence-v2
"1852140aef0a9d7900bae77e88b08a5e0cce01e1"
[ "citation", "other" ]
[ "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-26T21:03:54Z
darribas/satellite_led_liverpool
readme-supported-paper-method-implementation
[ "ml-context-only", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
14
[ "data", "machine-learning", "paper", "remote-sensing", "reproducibility", "socio-economic-indicators" ]
2025-05-31T01:13:07Z
https://github.com/darribas/satellite_led_liverpool
null
[ "general-ml" ]
[ "paper-implementation" ]
[ "description", "github-topics", "paper-reference", "query-match", "repository-metadata" ]
[ "general.arxiv" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-04-13T17:44:53Z
Hybrid Code Networks https://arxiv.org/abs/1702.03274
[ "general-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
21
88,191,165
null
Python
null
[ "paper-implementation" ]
johndpope/hcn
[ "description", "github-topics", "paper-reference", "query-match", "repository-metadata" ]
1
2026-09-24T16:52:35Z
[]
2017-04-13T17:43:32Z
[ "general.arxiv" ]
d555b918e3ddc52656ec9434dee20fd5c5c0d678
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"d555b918e3ddc52656ec9434dee20fd5c5c0d678"
[ "installation", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-27T17:16:20Z
johndpope/hcn
readme-supported-paper-method-implementation
[ "method-contribution", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
80
[ "bag-of-words", "dialog", "entitytracker", "gensim", "hybrid-code-networks", "lstm", "rnn", "tensorflow", "utterance" ]
2025-01-17T13:05:41Z
https://github.com/johndpope/hcn
null
[ "classical-ml", "probabilistic-ml", "representation-learning" ]
[ "gaussian-process", "kernel-methods", "metric-learning", "representation-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "general.gaussian-process", "general.metric-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-04-14T11:49:57Z
Code for the paper "Gaussian Process Classification as Metric Learning for Forensic Writer Identification", published at DAS 2018
[ "classical-ml", "probabilistic-ml" ]
[ "classifier" ]
ml_related_text
gh-ml-relevance-v1
2026-09-26T10:28:13Z
false
0
88,263,462
null
Python
MIT
[ "gaussian-process", "kernel-methods" ]
fredrikwahlberg/das2018
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
2
2026-09-27T16:09:59Z
[]
2018-05-03T13:28:29Z
[ "general.gaussian-process" ]
2a92a732848d696c61433274e2dd39196d2519ed
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"2a92a732848d696c61433274e2dd39196d2519ed"
[]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-27T17:16:20Z
fredrikwahlberg/das2018
readme-supported-paper-method-implementation
[ "method-contribution", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
2
[ "document-analysis", "gaussian-processes", "multi-class-classification", "unsupervised-feature-learning", "writer-indentification" ]
2020-09-24T16:59:04Z
https://github.com/fredrikwahlberg/das2018
null
[ "general-ml" ]
[ "paper-implementation" ]
[ "description", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
[ "general.arxiv" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-04-17T10:00:17Z
Source code for the EMNLP'17 paper "Deep Joint Entity Disambiguation with Local Neural Attention", https://arxiv.org/abs/1704.04920
[ "general-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T16:52:35Z
false
50
88,495,676
null
Lua
Apache-2.0
[ "paper-implementation" ]
dalab/deep-ed
[ "description", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
1
2026-09-24T16:52:35Z
[]
2018-01-12T12:11:19Z
[ "general.arxiv" ]
856f857497ed2a61103bd5346cc9f43cbf0f9baf
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"856f857497ed2a61103bd5346cc9f43cbf0f9baf"
[ "other" ]
[ "method-contribution", "ml-method-context", "model-training-artifact", "paper-code-relationship", "paper-reference" ]
ok
2026-09-27T17:16:20Z
dalab/deep-ed
readme-supported-paper-method-implementation
[ "method-contribution", "ml-context-only", "ml-method-context", "model-training-artifact", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
222
[]
2026-06-18T07:56:43Z
https://github.com/dalab/deep-ed
null
[ "multimodal", "multimodal-learning" ]
[ "convolutional-neural-network", "multimodal-learning", "neural-network" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "recall.name.multimodal-model" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-04-18T02:51:06Z
Code for paper "Musical Instrument Recognition in User-generated Videos using a Multimodal Convolutional Neural Network Architecture" by Olga Slizovskaia, Emilia Gomez, Gloria Haro. ICMR 2017
[ "multimodal", "multimodal-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T14:50:46Z
false
1
88,574,254
null
Python
GPL-3.0
[ "convolutional-neural-network", "multimodal-learning", "neural-network" ]
Veleslavia/ICMR2017
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-26T14:50:46Z
[]
2017-12-19T14:59:59Z
[ "recall.name.multimodal-model" ]
26bc02c9abc5f6be9bd8374ff5d183a273f2a08a
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"26bc02c9abc5f6be9bd8374ff5d183a273f2a08a"
[ "installation", "other", "usage" ]
[ "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-27T17:16:20Z
Veleslavia/ICMR2017
readme-supported-paper-method-implementation
[ "ml-context-only", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
8
[]
2022-12-18T12:23:11Z
https://github.com/Veleslavia/ICMR2017
null
[ "general-ml", "multimodal", "multimodal-learning", "reinforcement-learning", "robotics" ]
[ "model-based-reinforcement-learning", "multimodal-learning", "paper-implementation", "reinforcement-learning", "world-model" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.arxiv", "recall.name.multimodal-model", "rl.model-based" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-04-20T17:40:30Z
Code for paper "Learning Multimodal Transition Dynamics for Model-Based Reinforcement Learning".
[ "general-ml", "multimodal-learning", "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
7
88,893,609
null
Python
MIT
[ "paper-implementation", "reinforcement-learning" ]
tmoer/multimodal_varinf
[ "description", "license-metadata", "query-match", "repository-metadata" ]
5
2026-09-26T14:50:46Z
[]
2018-05-24T11:17:50Z
[ "general.arxiv" ]
041b494035a7fcc5e6bb26ea6d91b06e34abebe1
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"041b494035a7fcc5e6bb26ea6d91b06e34abebe1"
[ "citation", "method", "other" ]
[ "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-27T17:16:20Z
tmoer/multimodal_varinf
official-paper-method-implementation
[ "ml-method-context", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
34
[]
2026-09-13T06:40:15Z
https://github.com/tmoer/multimodal_varinf
null
[ "general-ml", "representation-learning" ]
[ "representation-learning", "semi-supervised-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.semi-supervised-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-04-24T19:06:03Z
Code for paper "Projected Estimators for Robust Semi-supervised Classification"
[ "general-ml", "representation-learning" ]
[ "classifier" ]
ml_related_text
gh-ml-relevance-v1
2026-09-26T14:50:46Z
false
1
89,277,125
null
TeX
null
[ "representation-learning", "semi-supervised-learning" ]
jkrijthe/ProjectedEstimators
[ "description", "query-match", "repository-metadata" ]
1
2026-09-26T14:50:46Z
[]
2017-04-24T19:10:41Z
[ "general.semi-supervised-learning" ]
f9bbdaa45824478abc3fa192945a07d614fa3304
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"f9bbdaa45824478abc3fa192945a07d614fa3304"
[ "abstract", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-27T17:16:20Z
jkrijthe/ProjectedEstimators
readme-supported-paper-method-implementation
[ "method-contribution", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
2
[]
2018-05-30T03:41:33Z
https://github.com/jkrijthe/ProjectedEstimators
null
[ "reinforcement-learning" ]
[ "deep-reinforcement-learning", "policy-learning", "reinforcement-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "rl.deep" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-05-28T13:14:59Z
Official implementation for the paper: "Shallow Updates for Deep Reinforcement Learning"
[ "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
5
92,662,121
null
Lua
null
[ "deep-reinforcement-learning", "policy-learning", "reinforcement-learning" ]
Shallow-Updates-for-Deep-RL/Shallow_Updates_for_Deep_RL
[ "description", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2017-11-02T18:52:18Z
[ "rl.deep" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
18
[]
2024-01-04T16:14:26Z
https://github.com/Shallow-Updates-for-Deep-RL/Shallow_Updates_for_Deep_RL
null
[ "recommender-systems" ]
[ "collaborative-filtering", "neural-network" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "applied.collaborative-filtering" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-05-29T17:57:45Z
Code for paper "On Sampling Strategies for Neural Network-based Collaborative Filtering"
[ "recommender-systems" ]
[ "deep-learning", "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
17
92,763,174
null
Python
MIT
[ "collaborative-filtering", "neural-network" ]
chentingpc/NNCF
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
4
2026-09-29T17:30:08Z
[]
2017-10-01T17:11:52Z
[ "applied.collaborative-filtering" ]
33a5e2fa64c5f43b025f85d81baac5cf605955f0
2026-09-28T20:05:39Z
gh-ml-readme-evidence-v2
"33a5e2fa64c5f43b025f85d81baac5cf605955f0"
[ "other" ]
[ "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-28T20:05:39Z
chentingpc/NNCF
readme-supported-paper-method-implementation
[ "ml-context-only", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
39
[ "collaborative-filtering", "deep-learning", "neural-networks", "recommender-system", "sgd" ]
2025-01-20T10:37:11Z
https://github.com/chentingpc/NNCF
null
[ "imitation-learning", "reinforcement-learning" ]
[ "imitation-learning", "inverse-reinforcement-learning", "reinforcement-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "rl.inverse" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-06-29T22:47:01Z
Implementations of Inverse Reinforcement Learning and new algorithms
[ "imitation-learning", "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
3
95,826,414
null
Python
null
[ "imitation-learning", "inverse-reinforcement-learning", "reinforcement-learning" ]
siddharthanpr/irl
[ "description", "query-match", "repository-metadata" ]
3
2026-09-25T16:01:25Z
[]
2017-06-29T22:55:35Z
[ "rl.inverse" ]
2ed058c5809522125cd20d9862ddeda5dc5bf65d
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
"2ed058c5809522125cd20d9862ddeda5dc5bf65d"
[ "other" ]
[ "ml-method-context" ]
ok
2026-09-26T15:53:33Z
siddharthanpr/irl
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-context", "ml-method-cue" ]
include
ml-contribution-v5
8
[]
2023-04-19T19:16:30Z
https://github.com/siddharthanpr/irl
null
[ "deep-learning", "embodied-ai", "multimodal", "representation-learning", "robotics" ]
[ "continual-learning", "lifelong-learning", "multi-task-learning", "robot-foundation-model", "robot-learning", "transfer-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "general.continual-learning", "general.multi-task-learning", "robotics.robot-foundation-model" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-07-13T21:07:15Z
Numenta published papers code and data
[ "embodied-ai", "multimodal", "robotics" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T15:29:21Z
false
94
97,165,046
null
Jupyter Notebook
AGPL-3.0
[ "robot-foundation-model", "robot-learning" ]
numenta/htmpapers
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
8
2026-09-29T17:30:08Z
[]
2022-03-30T23:59:13Z
[ "robotics.robot-foundation-model" ]
8d5c5efb658a9955d0b5a51303a706cda36321b4
2026-09-28T20:05:39Z
gh-ml-readme-evidence-v2
"8d5c5efb658a9955d0b5a51303a706cda36321b4"
[ "dataset", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-28T20:05:39Z
numenta/htmpapers
readme-supported-paper-method-implementation
[ "method-contribution", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
341
[ "htm", "neocortex", "numenta", "paper", "theory" ]
2026-07-20T07:37:32Z
https://github.com/numenta/htmpapers
null
[ "natural-language-processing", "reinforcement-learning" ]
[ "information-extraction", "reinforcement-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "nlp.information-extraction" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-07-20T16:27:23Z
Code for the paper 'Speeding up Reinforcement Learning-based Information Extraction Training using Asynchronous Methods' - EMNLP 2017
[ "natural-language-processing", "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
4
97,856,513
null
Roff
MIT
[ "information-extraction", "reinforcement-learning" ]
adi-sharma/RLIE_A3C
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2018-09-10T09:17:07Z
[ "nlp.information-extraction" ]
46e10caf20928c362886819429c47d943971d5f1
2026-09-28T20:05:39Z
gh-ml-readme-evidence-v2
"46e10caf20928c362886819429c47d943971d5f1"
[ "citation", "method", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-28T20:05:39Z
adi-sharma/RLIE_A3C
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-method-context", "ml-method-cue", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
8
[]
2025-08-25T16:38:44Z
https://github.com/adi-sharma/RLIE_A3C
null
[ "graph-learning" ]
[ "convolutional-neural-network", "graph-classification", "graph-representation-learning", "neural-network" ]
[ "description", "github-topics", "paper-reference", "query-match", "repository-metadata" ]
[ "graph.graph-classification" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-07-28T14:29:49Z
Code and data for the paper 'Classifying Graphs as Images with Convolutional Neural Networks' (new title: 'Graph Classification with 2D Convolutional Neural Networks')
[ "graph-learning" ]
[ "deep-learning", "neural-network", "representation-learning", "classifier", "embedding" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
23
98,657,591
https://arxiv.org/abs/1708.02218
Python
null
[ "convolutional-neural-network", "graph-classification", "graph-representation-learning", "neural-network" ]
Tixierae/graph_2D_CNN
[ "description", "github-topics", "paper-reference", "query-match", "repository-metadata" ]
4
2026-09-25T16:01:25Z
[]
2019-12-13T08:46:13Z
[ "graph.graph-classification" ]
3282c32987b3da87fa6d3e7bcd1c32dfc3b8c75e
2026-09-28T20:05:39Z
gh-ml-readme-evidence-v2
"3282c32987b3da87fa6d3e7bcd1c32dfc3b8c75e"
[ "citation", "installation", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-28T20:05:39Z
Tixierae/graph_2D_CNN
readme-supported-paper-method-implementation
[ "method-contribution", "ml-context-only", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
77
[ "2d-cnn", "artificial-intelligence", "classification", "convolutional-neural-networks", "deep-learning", "embeddings", "graph-2d-cnn", "graph-kernels", "graph-theory", "keras", "neural-networks", "representation-learning", "tensorflow" ]
2026-02-09T21:43:11Z
https://github.com/Tixierae/graph_2D_CNN
null
[ "deep-learning", "representation-learning" ]
[ "multi-task-learning", "transfer-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.multi-task-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-09-13T13:53:42Z
Code to replicate Experiments from the paper 'End-to-End Supervised Lobe Segmentation'
[ "deep-learning", "representation-learning" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-28T18:50:00Z
false
12
103,406,163
null
Python
null
[ "multi-task-learning", "transfer-learning" ]
filipetrocadoferreira/end2endlobesegmentation
[ "description", "query-match", "repository-metadata" ]
1
2026-09-28T18:50:00Z
[]
2018-03-29T10:39:29Z
[ "general.multi-task-learning" ]
182d6d00ded65e0a2f7b43a948f5989ab63ae25e
2026-09-29T18:47:14Z
gh-ml-readme-evidence-v2
"182d6d00ded65e0a2f7b43a948f5989ab63ae25e"
[ "method", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-29T18:47:14Z
filipetrocadoferreira/end2endlobesegmentation
readme-supported-paper-method-implementation
[ "method-contribution", "ml-method-context", "paper-and-code-cue", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
38
[]
2025-12-04T01:54:52Z
https://github.com/filipetrocadoferreira/end2endlobesegmentation
null
[ "computer-vision", "generative-ai", "generative-modeling" ]
[ "generative-modeling", "image-to-image-translation" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "vision.image-to-image-translation" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2017-11-27T01:43:01Z
StarGAN - Official PyTorch Implementation (CVPR 2018)
[ "computer-vision", "generative-ai", "generative-modeling" ]
[ "generative-model" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
953
112,133,243
null
Python
MIT
[ "generative-modeling", "image-to-image-translation" ]
yunjey/stargan
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
6
2026-09-28T18:50:00Z
[]
2021-01-23T15:09:58Z
[ "vision.image-to-image-translation" ]
bdd147fb2fff356e072dbae29550ea79cef44eb7
2026-09-25T22:08:11Z
gh-ml-readme-evidence-v1
"bdd147fb2fff356e072dbae29550ea79cef44eb7"
[ "citation", "other" ]
[ "method-contribution", "ml-method-context", "official-implementation-claim", "paper-code-relationship", "paper-reference" ]
ok
2026-09-25T22:08:11Z
yunjey/stargan
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-method-context", "official-implementation-claim", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
5,295
[ "cvpr2018", "generative-models", "image-to-image-translation", "pytorch", "stargan" ]
2026-09-22T20:17:59Z
https://github.com/yunjey/stargan
null
[ "reinforcement-learning" ]
[ "deep-reinforcement-learning", "policy-learning", "reinforcement-learning" ]
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
[ "rl.deep" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-01-03T12:53:15Z
Source code for paper Classification with Costly Features using Deep Reinforcement Learning.
[ "reinforcement-learning" ]
[ "reinforcement-learning", "classifier" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
24
116,136,597
https://arxiv.org/abs/1711.07364
Python
MIT
[ "deep-reinforcement-learning", "policy-learning", "reinforcement-learning" ]
jaromiru/cwcf
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2021-10-05T03:02:39Z
[ "rl.deep" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
58
[ "classification", "costly-features", "deep-reinforcement-learning" ]
2026-07-28T10:46:51Z
https://github.com/jaromiru/cwcf
null
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[ "distillation", "model-compression", "quantization" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "efficiency.model-compression", "efficiency.quantization" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-02-15T17:06:17Z
Implements quantized distillation. Code for our paper "Model compression via distillation and quantization"
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:27:52Z
false
76
121,656,522
null
Python
MIT
[ "distillation", "quantization" ]
antspy/quantized_distillation
[ "description", "license-metadata", "query-match", "repository-metadata" ]
7
2026-09-28T18:50:00Z
[]
2024-07-25T10:12:38Z
[ "efficiency.quantization" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
335
[]
2026-09-07T07:57:04Z
https://github.com/antspy/quantized_distillation
null
[ "deep-learning", "representation-learning" ]
[ "representation-learning", "self-supervised-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.self-supervised-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-02-15T20:02:58Z
Code for Paper: Self-supervised Learning of Motion Capture
[ "deep-learning", "representation-learning" ]
[ "self-supervised-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
13
121,676,445
null
Python
null
[ "representation-learning", "self-supervised-learning" ]
htung0101/3d_smpl
[ "description", "query-match", "repository-metadata" ]
6
2026-09-26T14:50:46Z
[]
2018-02-15T20:15:37Z
[ "general.self-supervised-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
90
[]
2025-04-05T22:43:41Z
https://github.com/htung0101/3d_smpl
null
[]
[ "meta-learning" ]
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
[]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-03-28T12:14:14Z
PyTorch code for CVPR 2018 paper: Learning to Compare: Relation Network for Few-Shot Learning (Few-Shot Learning part)
[]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
263
127,135,121
null
Python
MIT
[ "meta-learning" ]
floodsung/LearningToCompare_FSL
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
1
2026-09-26T20:57:28.489717Z
[]
2019-10-22T03:19:44Z
[]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
1,076
[ "few-shot-learning", "meta-learning" ]
2026-09-05T07:28:58Z
https://github.com/floodsung/LearningToCompare_FSL
{"candidate_evidence":["census-rule:liberal-ml-ai-v1","classifier-method:meta-learning"],"candidate_rule":"liberal-ml-ai-v1","discovery_source":"topic","queryless":true,"topic_names":["meta-learning"]}
[ "data-centric-ai", "machine-learning" ]
[ "active-learning", "convolutional-neural-network", "neural-network", "sample-selection" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.active-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-06-12T11:47:04Z
Source code for ICLR 2018 Paper: Active Learning for Convolutional Neural Networks: A Core-Set Approach
[ "data-centric-ai", "machine-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
47
137,063,829
null
Python
MIT
[ "active-learning", "convolutional-neural-network", "neural-network", "sample-selection" ]
ozansener/active_learning_coreset
[ "description", "license-metadata", "query-match", "repository-metadata" ]
5
2026-09-28T18:50:00Z
[]
2018-10-23T13:57:25Z
[ "general.active-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-context-only", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
282
[]
2026-08-24T20:31:26Z
https://github.com/ozansener/active_learning_coreset
null
[ "deep-learning", "graph-learning", "natural-language-processing" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-06-12T14:15:16Z
Accompanying code for our COLING 2018 paper "Modeling Semantics with Gated Graph Neural Networks for Knowledge Base Question Answering"
[ "deep-learning", "graph-learning", "natural-language-processing" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
26
137,081,892
null
Python
Apache-2.0
[ "graph-neural-network", "message-passing", "neural-network" ]
UKPLab/coling2018-graph-neural-networks-question-answering
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2020-02-12T13:06:17Z
[ "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
175
[]
2026-04-27T14:30:42Z
https://github.com/UKPLab/coling2018-graph-neural-networks-question-answering
null
[ "reinforcement-learning" ]
[ "deep-reinforcement-learning", "policy-learning", "reinforcement-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "rl.deep" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-06-26T11:53:36Z
Official Tensorflow implementation of drl-RPN: Deep Reinforcement Learning of Region Proposal Networks (CVPR 2018 paper)
[ "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
21
138,731,040
null
Jupyter Notebook
MIT
[ "deep-reinforcement-learning", "policy-learning", "reinforcement-learning" ]
aleksispi/drl-rpn-tf
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2024-02-22T13:54:19Z
[ "rl.deep" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
78
[]
2025-10-08T12:41:18Z
https://github.com/aleksispi/drl-rpn-tf
null
[ "natural-language-processing", "reinforcement-learning" ]
[ "natural-language-inference", "reinforcement-learning", "textual-entailment" ]
[ "description", "query-match", "repository-metadata" ]
[ "nlp-natural-language-inference" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-08-05T08:39:57Z
Code for ACL 2018 paper "Discourse Marker Augmented Network with Reinforcement Learning for Natural Language Inference".
[ "natural-language-processing", "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T14:50:46Z
false
4
143,596,688
null
Python
null
[ "natural-language-inference", "reinforcement-learning", "textual-entailment" ]
ZJULearning/DMP
[ "description", "query-match", "repository-metadata" ]
2
2026-09-27T16:09:59Z
[]
2018-08-05T09:21:45Z
[ "nlp-natural-language-inference" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
17
[]
2022-02-24T07:13:19Z
https://github.com/ZJULearning/DMP
null
[ "automl", "general-ml" ]
[ "few-shot-learning", "meta-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.meta-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-08-18T04:36:49Z
code for our IJCAI 2018 paper : "Lifelong Domain Word Embedding via Meta-Learning"
[ "automl", "general-ml" ]
[ "embedding" ]
ml_related_text
gh-ml-relevance-v1
2026-09-28T18:50:00Z
false
4
145,190,284
null
C
MIT
[ "few-shot-learning", "meta-learning" ]
howardhsu/L-DEM
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-28T18:50:00Z
[]
2019-04-02T21:40:48Z
[ "general.meta-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
11
[]
2020-02-18T03:55:57Z
https://github.com/howardhsu/L-DEM
null
[ "deep-learning", "representation-learning" ]
[ "continual-learning", "lifelong-learning", "neural-network" ]
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
[ "general.continual-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-10-12T15:47:58Z
This code refers to all experiments in our paper "Autonomous Deep Learning: Continual Learning Approach for Dynamic Environments"
[ "deep-learning", "representation-learning" ]
[ "deep-learning", "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
1
152,773,989
https://epubs.siam.org/doi/abs/10.1137/1.9781611975673.75
MATLAB
NOASSERTION
[ "continual-learning", "lifelong-learning", "neural-network" ]
andriash001/ADL
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2019-11-23T14:54:51Z
[ "general.continual-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-context-only", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
2
[ "autonomous-deep-learning", "continual-learning", "data-streams", "deep-learning", "deep-neural-networks" ]
2022-01-25T09:12:32Z
https://github.com/andriash001/ADL
null
[ "classical-ml", "tabular-and-structured-data", "tabular-ml" ]
[ "ensemble-learning", "gradient-boosting" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "general.gradient-boosting" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-10-26T13:12:25Z
This is the official clone for the implementation of the NIPS18 paper Multi-Layered Gradient Boosting Decision Trees (mGBDT) .
[ "classical-ml", "tabular-and-structured-data", "tabular-ml" ]
[ "representation-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
25
154,836,868
null
Python
null
[ "ensemble-learning", "gradient-boosting" ]
kingfengji/mGBDT
[ "description", "github-topics", "query-match", "repository-metadata" ]
5
2026-09-27T16:09:59Z
[]
2018-11-19T07:28:10Z
[ "general.gradient-boosting" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
103
[ "gbdt", "gradient-boosting-decision-trees", "mgbdt", "representation-learning", "target-propagation" ]
2026-07-08T19:07:27Z
https://github.com/kingfengji/mGBDT
null
[ "deep-learning", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-11-05T15:16:38Z
Source code for our AAAI paper "Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks".
[ "deep-learning", "graph-learning" ]
[ "deep-learning", "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
44
156,237,794
null
C++
null
[ "graph-neural-network", "message-passing", "neural-network" ]
chrsmrrs/k-gnn
[ "description", "github-topics", "query-match", "repository-metadata" ]
2
2026-09-27T16:09:59Z
[]
2022-03-22T12:39:40Z
[ "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
191
[ "deep-learning", "end-to-end", "graph-algorithms", "graph-neural-networks", "graphs", "higher-order", "pytorch", "weisfeier-leman", "weisfeiler-lehman" ]
2026-04-08T10:37:08Z
https://github.com/chrsmrrs/k-gnn
null
[ "classical-ml", "computer-vision", "efficient-ml", "model-compression", "representation-learning" ]
[ "distillation", "knowledge-distillation", "metric-learning", "neural-network", "representation-learning" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "efficiency.knowledge-distillation", "general.metric-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-11-24T13:25:12Z
Official pytorch Implementation of Relational Knowledge Distillation, CVPR 2019
[ "classical-ml", "computer-vision", "representation-learning" ]
[ "deep-learning", "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
51
158,938,672
null
Python
null
[ "distillation", "knowledge-distillation", "metric-learning", "neural-network", "representation-learning" ]
lenscloth/RKD
[ "description", "github-topics", "query-match", "repository-metadata" ]
9
2026-09-28T18:50:00Z
[]
2021-05-17T04:00:24Z
[ "general.metric-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue" ]
include
ml-contribution-v5
420
[ "computer-vision", "deep-learning", "deep-neural-networks", "knowledge-distillation", "metric-learning" ]
2026-09-09T01:48:58Z
https://github.com/lenscloth/RKD
null
[ "distributed-ml", "privacy-and-federated-learning" ]
[ "collaborative-learning", "federated-learning" ]
[ "description", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
[ "trust.federated-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2018-12-12T20:57:32Z
Source code for paper "How to Backdoor Federated Learning" (https://arxiv.org/abs/1807.00459)
[ "distributed-ml", "privacy-and-federated-learning" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
63
161,544,036
null
Python
MIT
[ "collaborative-learning", "federated-learning" ]
ebagdasa/backdoor_federated_learning
[ "description", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
5
2026-09-27T16:09:59Z
[]
2024-07-25T10:14:50Z
[ "trust.federated-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
316
[]
2026-09-21T05:09:04Z
https://github.com/ebagdasa/backdoor_federated_learning
null
[ "reinforcement-learning" ]
[ "distributional-reinforcement-learning", "reinforcement-learning", "value-based-reinforcement-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "rl.distributional" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-01-03T05:26:26Z
Rlee is a research framework built on top of PyTorch 1.0 for fast prototyping of novel reinforcement learning algorithms.
[ "reinforcement-learning" ]
[ "deep-learning", "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T10:28:13Z
false
0
163,927,063
https://www.endtoend.ai
Python
MIT
[ "distributional-reinforcement-learning", "reinforcement-learning", "value-based-reinforcement-learning" ]
seungjaeryanlee/rlee
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-26T10:28:13Z
[]
2023-07-06T21:31:51Z
[ "rl.distributional" ]
43eb8512d46971fe41345d02069e6d31ecd8a6ad
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
"43eb8512d46971fe41345d02069e6d31ecd8a6ad"
[ "other" ]
[ "ml-method-context", "paper-reference" ]
ok
2026-09-26T15:53:33Z
seungjaeryanlee/rlee
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-context", "ml-method-cue", "paper-reference" ]
include
ml-contribution-v5
2
[ "deep-learning", "deep-reinforcement-learning", "python", "pytorch", "reinforcement-learning" ]
2024-01-09T11:17:53Z
https://github.com/seungjaeryanlee/rlee
null
[ "automl", "general-ml" ]
[ "few-shot-learning", "meta-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.meta-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-01-28T15:00:04Z
Official code for the paper titled "Meta Learning Deep Visual Words for Fast Video Object Segmentation"
[ "automl", "general-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-28T18:50:00Z
false
3
167,980,836
null
Python
MIT
[ "few-shot-learning", "meta-learning" ]
harkiratbehl/MetaVOS
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-28T18:50:00Z
[]
2020-04-13T10:58:10Z
[ "general.meta-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
23
[]
2024-04-09T09:03:15Z
https://github.com/harkiratbehl/MetaVOS
null
[ "reinforcement-learning" ]
[ "batch-reinforcement-learning", "offline-reinforcement-learning", "reinforcement-learning" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "rl.offline" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-02-02T09:18:31Z
[AAAI 2022] The official implementation of "DeepThermal: Combustion Optimization for Thermal Power Generating Units Using Offline Reinforcement Learning"
[ "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
2
168,814,886
null
Python
null
[ "batch-reinforcement-learning", "offline-reinforcement-learning", "reinforcement-learning" ]
ryanxhr/DeepThermal
[ "description", "github-topics", "query-match", "repository-metadata" ]
4
2026-09-25T16:01:25Z
[]
2022-07-21T07:40:15Z
[ "rl.offline" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue" ]
include
ml-contribution-v5
21
[ "model-based-reinforcement-learning", "offline-reinforcement-learning", "tensorflow" ]
2026-07-20T07:41:57Z
https://github.com/ryanxhr/DeepThermal
null
[ "linguistics", "natural-language-processing" ]
[ "language-modeling" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "recall.computational-linguistics" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-02-24T20:52:37Z
code for our NAACL 2019 paper: "BERT Post-Training for Review Reading Comprehension and Aspect-based Sentiment Analysis"
[ "linguistics", "natural-language-processing" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-25T16:01:25Z
false
110
172,388,988
null
Python
Apache-2.0
[ "language-modeling" ]
howardhsu/BERT-for-RRC-ABSA
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-25T16:01:25Z
[]
2021-02-05T05:58:43Z
[ "recall.computational-linguistics" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
461
[ "bert", "reading-comprehension", "sentiment-analysis" ]
2026-09-08T02:37:55Z
https://github.com/howardhsu/BERT-for-RRC-ABSA
null
[ "computer-vision", "interpretability-and-safety", "trustworthy-ml" ]
[ "explainable-ai", "interpretability" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "trust.explainable-ai" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-03-11T12:28:20Z
Official PyTorch implementation of "Visualizing the Decision-making Process in Deep Neural Decision Forest", CVPR 2019 Workshops on Explainable AI
[ "computer-vision", "interpretability-and-safety", "trustworthy-ml" ]
[ "deep-learning", "machine-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
19
174,995,387
null
Python
MIT
[ "explainable-ai", "interpretability" ]
Nicholasli1995/VisualizingNDF
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
4
2026-09-26T14:50:46Z
[]
2022-03-12T06:32:59Z
[ "trust.explainable-ai" ]
b74f875c209061e18922f285e544a37f7b731b34
2026-09-25T22:08:11Z
gh-ml-readme-evidence-v1
"b74f875c209061e18922f285e544a37f7b731b34"
[ "citation", "other", "results", "usage" ]
[ "method-contribution", "ml-method-context", "official-implementation-claim", "paper-reference" ]
ok
2026-09-25T22:08:11Z
Nicholasli1995/VisualizingNDF
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-context-only", "ml-method-context", "official-implementation-claim", "paper-reference" ]
include
ml-contribution-v5
73
[ "age-estimation", "cifar10", "computer-vision", "deep-learning", "imageclassification", "machine-learning", "mnist", "visualization" ]
2026-03-26T17:25:26Z
https://github.com/Nicholasli1995/VisualizingNDF
null
[ "generative-ai", "language" ]
[ "fine-tuning", "parameter-efficient-fine-tuning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "llm.finetuning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-03-25T02:05:03Z
Code for paper Fine-tune BERT for Extractive Summarization
[ "generative-ai", "language" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
410
177,497,186
null
Python
Apache-2.0
[ "fine-tuning", "parameter-efficient-fine-tuning" ]
nlpyang/BertSum
[ "description", "license-metadata", "query-match", "repository-metadata" ]
5
2026-09-28T18:50:00Z
[]
2022-01-11T07:58:23Z
[ "llm.finetuning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
1,504
[]
2026-09-24T17:44:57Z
https://github.com/nlpyang/BertSum
null
[ "automl", "general-ml" ]
[ "few-shot-learning", "meta-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.meta-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-04-15T06:50:49Z
Code for paper "Learning to Guide: Guidance Law Based on Deep Meta-learning and Model Predictive Path Integral Control"
[ "automl", "general-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-28T18:50:00Z
false
2
181,428,954
null
Python
MIT
[ "few-shot-learning", "meta-learning" ]
tccliangchen/deep_meta-learning_guidance_law
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-28T18:50:00Z
[]
2019-05-26T10:52:29Z
[ "general.meta-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
12
[]
2026-09-26T16:45:43Z
https://github.com/tccliangchen/deep_meta-learning_guidance_law
null
[ "distributed-ml", "privacy-and-federated-learning" ]
[ "collaborative-learning", "federated-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "trust.federated-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-05-09T02:34:51Z
Code for paper "Interpret Federated Learning with Shapley Values"
[ "distributed-ml", "privacy-and-federated-learning" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
5
185,712,679
null
Jupyter Notebook
Apache-2.0
[ "collaborative-learning", "federated-learning" ]
crownpku/federated_shap
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2019-05-18T01:58:52Z
[ "trust.federated-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
39
[]
2026-08-10T20:20:16Z
https://github.com/crownpku/federated_shap
null
[ "reinforcement-learning" ]
[ "meta-learning", "reinforcement-learning" ]
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
[]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-05-17T02:30:00Z
Implementation of our paper "Meta Reinforcement Learning with Task Embedding and Shared Policy"
[ "reinforcement-learning" ]
[ "machine-learning", "reinforcement-learning", "embedding" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
7
187,133,156
null
Python
NOASSERTION
[ "meta-learning", "reinforcement-learning" ]
llan-ml/tesp
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
1
2026-09-26T20:57:28.489717Z
[]
2019-05-17T11:21:17Z
[]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
35
[ "meta-learning", "meta-reinforcement", "meta-rl", "reinforcement-learning", "tesp" ]
2025-11-16T07:43:02Z
https://github.com/llan-ml/tesp
{"candidate_evidence":["census-rule:liberal-ml-ai-v1","classifier-domain:reinforcement-learning","classifier-method:meta-learning","classifier-method:reinforcement-learning","generic-ml-ai-term"],"candidate_rule":"liberal-ml-ai-v1","discovery_source":"topic","queryless":true,"topic_names":["meta-learning"]}
[]
[ "meta-learning" ]
[ "description", "github-topics", "repository-metadata" ]
[]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-05-17T04:22:53Z
The code for paper "CANet: Class-Agnostic Segmentation Networks with Iterative Refinement and Attentive Few-Shot Learning"
[]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
33
187,146,048
null
Python
null
[ "meta-learning" ]
icoz69/CaNet
[ "description", "github-topics", "repository-metadata" ]
1
2026-09-26T20:57:28.489717Z
[]
2020-06-06T10:55:56Z
[]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
198
[ "cvpr2019", "few-shot-learning", "meta-learning", "segmentation" ]
2026-06-16T07:14:08Z
https://github.com/icoz69/CaNet
{"candidate_evidence":["census-rule:liberal-ml-ai-v1","classifier-method:meta-learning"],"candidate_rule":"liberal-ml-ai-v1","discovery_source":"topic","queryless":true,"topic_names":["meta-learning"]}
[ "3d-vision", "computer-vision", "efficient-ml", "model-compression" ]
[ "3d-perception", "depth-estimation", "distillation", "knowledge-distillation" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "efficiency.knowledge-distillation", "vision.depth-estimation" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-05-27T05:31:22Z
The official code for the paper 'Structured Knowledge Distillation for Semantic Segmentation'. (CVPR 2019 ORAL) and extension to other tasks.
[ "3d-vision", "computer-vision" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
106
188,776,740
null
Python
BSD-2-Clause
[ "3d-perception", "depth-estimation", "distillation", "knowledge-distillation" ]
irfanICMLL/structure_knowledge_distillation
[ "description", "license-metadata", "query-match", "repository-metadata" ]
8
2026-09-28T18:50:00Z
[]
2020-04-20T06:49:03Z
[ "vision.depth-estimation" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
739
[]
2026-08-20T21:02:24Z
https://github.com/irfanICMLL/structure_knowledge_distillation
null
[ "automl", "general-ml" ]
[ "few-shot-learning", "meta-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.meta-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-05-28T12:48:11Z
Code for paper "Learning to Guide: Guidance Law Based on Deep Meta-learning and Model Predictive Path Integral Control"
[ "automl", "general-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-28T18:50:00Z
false
5
189,027,100
null
Python
MIT
[ "few-shot-learning", "meta-learning" ]
BenjyWP/deep_meta-learning_guidance_law
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-28T18:50:00Z
[]
2019-05-26T10:52:29Z
[ "general.meta-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
19
[]
2026-05-17T03:17:56Z
https://github.com/BenjyWP/deep_meta-learning_guidance_law
null
[ "multi-agent-systems", "reinforcement-learning" ]
[ "centralized-training", "multi-agent-reinforcement-learning", "reinforcement-learning" ]
[ "description", "query-match", "repository-metadata" ]
[ "rl.multiagent" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-05-29T02:29:00Z
Source code for paper:Multi-agent reinforcement learning for liquidation strategy analysis
[ "multi-agent-systems", "reinforcement-learning" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
16
189,136,065
null
Jupyter Notebook
null
[ "centralized-training", "multi-agent-reinforcement-learning", "reinforcement-learning" ]
WenhangBao/Multi-Agent-RL-for-Liquidation
[ "description", "query-match", "repository-metadata" ]
4
2026-09-26T14:50:46Z
[]
2019-05-30T00:02:59Z
[ "rl.multiagent" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
60
[]
2026-06-05T11:59:57Z
https://github.com/WenhangBao/Multi-Agent-RL-for-Liquidation
null
[ "computer-vision", "generative-ai", "generative-modeling", "multimodal", "video" ]
[ "generative-modeling", "text-to-video", "video-generation" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "multimodal.text-to-video", "vision.video-generation" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-05-31T17:00:00Z
Code for our IJCAI 2019 paper entitled "Conditional GAN with Discriminative Filter Generation for Text-to-Video Synthesis"
[ "computer-vision", "generative-ai", "generative-modeling" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-26T14:50:46Z
false
3
189,629,698
null
Python
BSD-2-Clause
[ "generative-modeling", "video-generation" ]
minrq/CGAN_Text2Video
[ "description", "license-metadata", "query-match", "repository-metadata" ]
2
2026-09-28T18:50:00Z
[]
2022-03-29T15:31:48Z
[ "vision.video-generation" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
14
[]
2024-01-02T17:14:26Z
https://github.com/minrq/CGAN_Text2Video
null
[ "generative-modeling" ]
[ "transformer" ]
[ "description", "github-topics", "repository-metadata" ]
[]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-06-02T15:34:38Z
New Transformer network-based GAN for video generation.
[ "generative-modeling" ]
[ "transformer" ]
ml_related_text
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
1
189,863,981
null
Jupyter Notebook
null
[ "transformer" ]
Nilanshrajput/Video_Generation_Transformer
[ "description", "github-topics", "repository-metadata" ]
1
2026-09-26T20:57:28.489717Z
[]
2020-06-01T05:57:24Z
[]
null
2026-09-26T21:03:54Z
gh-ml-readme-evidence-v2
null
[]
[]
missing
2026-09-26T21:03:54Z
Nilanshrajput/Video_Generation_Transformer
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-cue" ]
include
ml-contribution-v5
3
[ "gan", "pytorch", "singan", "video-generation" ]
2023-08-28T11:07:04Z
https://github.com/Nilanshrajput/Video_Generation_Transformer
{"candidate_evidence":["census-rule:liberal-ml-ai-v1","classifier-domain:generative-modeling","classifier-method:transformer","generic-ml-ai-term"],"candidate_rule":"liberal-ml-ai-v1","discovery_source":"topic","queryless":true,"topic_names":["video-generation"]}
[ "natural-language-processing" ]
[ "abstractive-summarization", "coreference-resolution", "discourse-understanding", "text-summarization" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "nlp.coreference-resolution", "nlp.text-summarization" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-06-17T20:39:02Z
Code for paper "Discourse-Aware Neural Extractive Text Summarization" (ACL20)
[ "natural-language-processing" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
30
192,415,533
null
Python
MIT
[ "abstractive-summarization", "coreference-resolution", "discourse-understanding", "text-summarization" ]
jiacheng-xu/DiscoBERT
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
5
2026-09-27T16:09:59Z
[]
2020-04-25T03:44:47Z
[ "nlp.coreference-resolution", "nlp.text-summarization" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
166
[ "acl2020", "bert-model", "microsoft-dynamics-365", "natural-language-processing", "text-summarization" ]
2026-03-05T05:06:13Z
https://github.com/jiacheng-xu/DiscoBERT
null
[ "computer-vision", "generative-ai" ]
[ "generative-modeling", "image-to-image-translation" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "vision.image-to-image-translation" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-07-26T00:33:54Z
Official Tensorflow implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation (ICLR 2020)
[ "computer-vision", "generative-ai" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
1,007
198,919,091
null
Python
MIT
[ "generative-modeling", "image-to-image-translation" ]
taki0112/UGATIT
[ "description", "license-metadata", "query-match", "repository-metadata" ]
6
2026-09-28T18:50:00Z
[]
2021-05-20T03:23:05Z
[ "vision.image-to-image-translation" ]
8705566f45d96a473d27a6b38ac2108592220fa3
2026-09-25T22:08:11Z
gh-ml-readme-evidence-v1
"8705566f45d96a473d27a6b38ac2108592220fa3"
[ "citation", "dataset", "installation", "method", "other", "usage" ]
[ "method-contribution", "ml-method-context", "official-implementation-claim", "paper-code-relationship", "paper-reference" ]
ok
2026-09-25T22:08:11Z
taki0112/UGATIT
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-method-context", "official-implementation-claim", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
6,113
[]
2026-09-17T05:00:39Z
https://github.com/taki0112/UGATIT
null
[ "computer-vision", "deep-learning", "graph-learning", "science-and-engineering" ]
[ "graph-neural-network", "graph-representation-learning", "message-passing", "neural-network" ]
[ "description", "query-match", "repository-metadata" ]
[ "graph.gnn-description", "graph.representation-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-07-29T06:36:04Z
Many underlying relationships among data in several areas of science and engineering, e.g., computer vision, molecular chemistry, molecular biology, pattern recognition, and data mining, can be represented in terms of graphs. In this paper ,a new neural network model has propose which called graph neural network (GNN)
[ "computer-vision", "deep-learning", "graph-learning", "science-and-engineering" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
9
199,393,193
null
Jupyter Notebook
null
[ "graph-neural-network", "graph-representation-learning", "message-passing", "neural-network" ]
Erfaan-Rostami/Hypergraph-and-Graph-Neural-Network-HGNN-GNN--
[ "description", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2026-05-21T14:24:22Z
[ "graph.gnn-description", "graph.representation-learning" ]
2aa544589c2b902b837d2c56e9a1d57a5e7af014
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"2aa544589c2b902b837d2c56e9a1d57a5e7af014"
[ "other" ]
[ "ml-method-context", "paper-reference", "survey-cue" ]
ok
2026-09-27T17:16:20Z
Erfaan-Rostami/Hypergraph-and-Graph-Neural-Network-HGNN-GNN--
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-context", "ml-method-cue", "paper-reference", "survey-cue" ]
include
ml-contribution-v5
58
[]
2026-06-02T14:37:23Z
https://github.com/Erfaan-Rostami/Hypergraph-and-Graph-Neural-Network-HGNN-GNN--
null
[ "computer-vision", "generative-ai" ]
[ "generative-modeling", "image-to-image-translation" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "vision.image-to-image-translation" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-07-29T07:44:56Z
Official PyTorch implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation
[ "computer-vision", "generative-ai" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
460
199,404,030
null
Python
MIT
[ "generative-modeling", "image-to-image-translation" ]
znxlwm/UGATIT-pytorch
[ "description", "license-metadata", "query-match", "repository-metadata" ]
6
2026-09-28T18:50:00Z
[]
2023-03-16T02:38:05Z
[ "vision.image-to-image-translation" ]
3ad2faea9f204973dc0c11ccee261656a3dd9b14
2026-09-25T22:08:11Z
gh-ml-readme-evidence-v1
"3ad2faea9f204973dc0c11ccee261656a3dd9b14"
[ "method", "other", "usage" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-25T22:08:11Z
znxlwm/UGATIT-pytorch
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
2,533
[]
2026-09-24T17:45:47Z
https://github.com/znxlwm/UGATIT-pytorch
null
[ "deep-learning", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-08-05T07:56:32Z
a novel DTA predition method using graph neural network
[ "deep-learning", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
42
200,609,566
null
Python
null
[ "graph-neural-network", "message-passing", "neural-network" ]
595693085/DGraphDTA
[ "description", "query-match", "repository-metadata" ]
7
2026-09-27T16:09:59Z
[]
2023-07-12T16:23:50Z
[ "graph.gnn-description" ]
1b63165b56276c5dd5805578ecb1ab77705e45e6
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
"1b63165b56276c5dd5805578ecb1ab77705e45e6"
[ "other" ]
[ "method-contribution", "ml-method-context" ]
ok
2026-09-26T15:53:33Z
595693085/DGraphDTA
specific-method-with-novelty-claim
[ "method-contribution", "method-tied-novelty-claim", "ml-method-context", "ml-method-cue" ]
include
ml-contribution-v5
77
[]
2026-07-10T04:59:29Z
https://github.com/595693085/DGraphDTA
null
[ "deep-learning", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-08-11T09:13:23Z
DrBC: A novel graph neural network approach to identify high Betweenness Centraliy (BC) nodes ( CIKM'19 )
[ "deep-learning", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
7
201,742,627
null
Python
MIT
[ "graph-neural-network", "message-passing", "neural-network" ]
fanchangjun/DrBC
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2020-12-20T13:46:08Z
[ "graph.gnn-description" ]
77539734a953855117f31d9a065580a9741bcd10
2026-09-27T17:16:20Z
gh-ml-readme-evidence-v2
"77539734a953855117f31d9a065580a9741bcd10"
[ "other" ]
[ "ml-method-context", "paper-reference" ]
ok
2026-09-27T17:16:20Z
fanchangjun/DrBC
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-context", "ml-method-cue", "paper-reference" ]
include
ml-contribution-v5
33
[]
2025-09-09T17:13:17Z
https://github.com/fanchangjun/DrBC
null
[ "computer-vision" ]
[ "distillation", "keypoint-detection", "knowledge-distillation", "pose-estimation" ]
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
[ "vision.pose-estimation" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-08-14T12:28:56Z
Official pytorch Code for CVPR2019 paper "Fast Human Pose Estimation" https://arxiv.org/abs/1811.05419
[ "computer-vision" ]
[ "deep-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
67
202,346,299
null
Cuda
MIT
[ "distillation", "keypoint-detection", "knowledge-distillation", "pose-estimation" ]
ilovepose/fast-human-pose-estimation.pytorch
[ "description", "github-topics", "license-metadata", "paper-reference", "query-match", "repository-metadata" ]
7
2026-09-28T18:50:00Z
[]
2022-09-16T07:27:38Z
[ "vision.pose-estimation" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
399
[ "coco-keypoints-detection", "deep-learning", "fast-pose-distillation", "human-pose-estimation", "knowledge-distillation", "mpii-dataset", "mscoco-keypoint" ]
2026-08-17T13:40:21Z
https://github.com/ilovepose/fast-human-pose-estimation.pytorch
null
[ "automl", "deep-learning", "general-ml", "graph-learning" ]
[ "few-shot-learning", "graph-neural-network", "message-passing", "meta-learning", "neural-network" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "general.meta-learning", "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-08-15T03:04:58Z
official PyTorch implementation of paper "Continual Meta-Learning with Bayesian Graph Neural Networks" (AAAI2020)
[ "automl", "deep-learning", "general-ml", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
17
202,464,515
null
Python
MIT
[ "few-shot-learning", "graph-neural-network", "message-passing", "meta-learning", "neural-network" ]
Luoyadan/BGNN-AAAI
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2020-06-18T09:26:50Z
[ "general.meta-learning", "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
65
[ "few-shot-learning", "graph-neural-networks", "meta-learning" ]
2026-09-27T07:45:00Z
https://github.com/Luoyadan/BGNN-AAAI
null
[ "deep-learning", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-09-06T06:19:32Z
Source code for paper "Knowledge-aware Heterogeneous Graph Neural Networks for Inferring Substitutable and Complementary Items"
[ "deep-learning", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
0
206,726,521
null
null
null
[ "graph-neural-network", "message-passing", "neural-network" ]
fanyubupt/KHGNN
[ "description", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2019-09-06T06:19:33Z
[ "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
0
[]
2019-09-06T06:19:32Z
https://github.com/fanyubupt/KHGNN
null
[]
[ "distillation", "knowledge-distillation" ]
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
[]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-09-10T03:26:42Z
Official PyTorch implementation of "A Comprehensive Overhaul of Feature Distillation" (ICCV 2019)
[]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
74
207,457,047
null
Python
MIT
[ "distillation", "knowledge-distillation" ]
clovaai/overhaul-distillation
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
1
2026-09-26T20:57:28.489717Z
[]
2020-06-23T09:33:49Z
[]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue" ]
include
ml-contribution-v5
421
[ "iccv2019", "knowledge-distillation", "knowledge-transfer", "network-compression", "teacher-student" ]
2026-09-19T05:38:46Z
https://github.com/clovaai/overhaul-distillation
{"candidate_evidence":["census-rule:liberal-ml-ai-v1","classifier-method:distillation","classifier-method:knowledge-distillation"],"candidate_rule":"liberal-ml-ai-v1","discovery_source":"topic","queryless":true,"topic_names":["knowledge-distillation"]}
[ "automl", "deep-learning", "efficient-ml", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-architecture-search", "neural-network" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "efficiency.neural-architecture-search", "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-09-16T17:20:34Z
Code for paper: Neural Architecture Search in Graph Neural Networks (BRACIS 2020)
[ "deep-learning", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-25T16:01:25Z
false
3
208,856,717
null
Jupyter Notebook
Apache-2.0
[ "graph-neural-network", "message-passing", "neural-network" ]
mhnnunes/nas_gnn
[ "description", "license-metadata", "query-match", "repository-metadata" ]
2
2026-09-27T16:09:59Z
[]
2023-07-06T21:27:58Z
[ "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
19
[]
2025-08-06T10:51:36Z
https://github.com/mhnnunes/nas_gnn
null
[ "forecasting", "time-series", "time-series-and-forecasting" ]
[ "forecasting", "sequence-modeling", "time-series-forecasting" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "applied.time-series-forecasting", "timeseries.forecasting" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-09-17T11:23:55Z
Code for our NeurIPS 2019 paper "Shape and Time Distortion Loss for Training Deep Time Series Forecasting Models"
[ "forecasting", "time-series", "time-series-and-forecasting" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
77
209,034,747
null
Python
NOASSERTION
[ "forecasting", "sequence-modeling" ]
vincent-leguen/DILATE
[ "description", "license-metadata", "query-match", "repository-metadata" ]
5
2026-09-27T16:09:59Z
[]
2020-10-14T12:33:08Z
[ "timeseries.forecasting" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
400
[]
2026-09-10T14:36:50Z
https://github.com/vincent-leguen/DILATE
null
[ "classical-ml" ]
[ "kernel-methods", "support-vector-machine" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.support-vector-machine" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-10-02T17:09:43Z
Code for paper: "Support Vector Machines, Wasserstein's distance and gradient-penalty GANs maximize a margin"
[ "classical-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
24
212,397,158
null
Python
MIT
[ "kernel-methods", "support-vector-machine" ]
AlexiaJM/MaximumMarginGANs
[ "description", "license-metadata", "query-match", "repository-metadata" ]
4
2026-09-29T17:30:08Z
[]
2020-03-12T14:52:28Z
[ "general.support-vector-machine" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
180
[]
2026-09-23T15:32:20Z
https://github.com/AlexiaJM/MaximumMarginGANs
null
[ "general-ml" ]
[ "novel-method", "random-forest" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.novel-method" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-10-05T10:16:30Z
The code implements a novel method for converting random forest into a single decision tree
[ "general-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
4
212,979,949
null
Python
null
[ "novel-method", "random-forest" ]
sagyome/forest_based_tree
[ "description", "query-match", "repository-metadata" ]
4
2026-09-25T16:01:25Z
[]
2020-01-24T21:55:20Z
[ "general.novel-method" ]
dcba2ed3ddb2180c4260f3dac349e1a398adce38
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
"dcba2ed3ddb2180c4260f3dac349e1a398adce38"
[ "other" ]
[]
ok
2026-09-26T15:53:33Z
sagyome/forest_based_tree
specific-method-with-novelty-claim
[ "contribution-language", "method-tied-novelty-claim", "ml-method-cue" ]
include
ml-contribution-v5
10
[]
2026-07-06T18:12:51Z
https://github.com/sagyome/forest_based_tree
null
[ "reinforcement-learning", "robotics" ]
[ "domain-randomization", "reinforcement-learning", "sim-to-real" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "robotics.domain-randomization" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-10-18T10:52:12Z
Code associated with our paper "Robust Domain Randomization for Reinforcement Learning"
[ "reinforcement-learning", "robotics" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
3
216,003,313
null
Python
MIT
[ "domain-randomization", "reinforcement-learning", "sim-to-real" ]
uncharted-technologies/robust-domain-randomization
[ "description", "license-metadata", "query-match", "repository-metadata" ]
3
2026-09-25T16:01:25Z
[]
2022-11-22T04:35:32Z
[ "robotics.domain-randomization" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
12
[]
2025-06-18T07:06:45Z
https://github.com/uncharted-technologies/robust-domain-randomization
null
[ "deep-learning", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-10-30T19:06:39Z
Official PyTorch Implementation for "GNN3DMOT: Graph Neural Network for 3D Multi-Object Tracking with 2D-3D Multi-Feature Learning", CVPR 2020
[ "deep-learning", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
6
218,602,535
http://www.xinshuoweng.com/
null
null
[ "graph-neural-network", "message-passing", "neural-network" ]
xinshuoweng/GNN3DMOT
[ "description", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2020-06-07T21:24:10Z
[ "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue" ]
include
ml-contribution-v5
81
[]
2026-07-04T07:46:11Z
https://github.com/xinshuoweng/GNN3DMOT
null
[ "deep-learning", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-11-12T11:20:13Z
The implementation of our ICDM 2019 paper "Relation Structure-Aware Heterogeneous Graph Neural Network" RSHN.
[ "deep-learning", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
4
221,200,352
null
Python
null
[ "graph-neural-network", "message-passing", "neural-network" ]
CheriseZhu/RSHN
[ "description", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2020-10-22T01:49:22Z
[ "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
18
[]
2023-11-17T07:45:01Z
https://github.com/CheriseZhu/RSHN
null
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[ "neural-network-pruning", "pruning", "structured-pruning" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "efficiency.pruning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-11-17T10:35:12Z
Pytorch implementation of our paper accepted by CVPR 2020 (Oral) -- HRank: Filter Pruning using High-Rank Feature Map
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
50
222,231,732
https://128.84.21.199/abs/2002.10179
Python
null
[ "neural-network-pruning", "pruning", "structured-pruning" ]
lmbxmu/HRank
[ "description", "github-topics", "query-match", "repository-metadata" ]
4
2026-09-29T17:30:08Z
[]
2021-02-11T16:48:26Z
[ "efficiency.pruning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
257
[ "acceleration", "compression", "pruning" ]
2026-01-30T12:38:52Z
https://github.com/lmbxmu/HRank
null
[ "computer-vision" ]
[ "self-supervised-learning" ]
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
[]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-11-22T06:08:58Z
Code for ICLR 2020 paper "VL-BERT: Pre-training of Generic Visual-Linguistic Representations".
[ "computer-vision" ]
[ "representation-learning", "self-supervised-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
109
223,335,609
null
Jupyter Notebook
MIT
[ "self-supervised-learning" ]
jackroos/VL-BERT
[ "description", "github-topics", "license-metadata", "repository-metadata" ]
1
2026-09-26T20:57:28.489717Z
[]
2023-05-22T22:33:35Z
[]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
740
[ "bert", "iclr2020", "pre-training", "pytorch", "representation-learning", "self-supervised-learning", "vision-and-language", "vl-bert" ]
2026-08-07T14:17:29Z
https://github.com/jackroos/VL-BERT
{"candidate_evidence":["census-rule:liberal-ml-ai-v1","classifier-domain:computer-vision","classifier-method:self-supervised-learning"],"candidate_rule":"liberal-ml-ai-v1","discovery_source":"topic","queryless":true,"topic_names":["self-supervised-learning"]}
[ "multimodal", "reinforcement-learning", "robotics" ]
[ "environment-modeling", "reinforcement-learning", "world-model" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "multimodal.world-models" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-11-26T12:08:03Z
A new version of world models using Echo-state networks and random weight-fixed CNNs
[ "multimodal", "reinforcement-learning", "robotics" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T14:50:46Z
false
1
224,183,627
null
Python
null
[ "environment-modeling", "reinforcement-learning", "world-model" ]
Shahdsaf/Semi-Supervised-World-Models
[ "description", "github-topics", "query-match", "repository-metadata" ]
2
2026-09-26T14:50:46Z
[]
2020-06-01T23:12:22Z
[ "multimodal.world-models" ]
9deeb85227d7695efc1e56d4f574db208e1f1b54
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
"9deeb85227d7695efc1e56d4f574db208e1f1b54"
[ "other", "results" ]
[ "ml-method-context", "model-training-artifact", "paper-reference" ]
ok
2026-09-26T15:53:33Z
Shahdsaf/Semi-Supervised-World-Models
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-context", "ml-method-cue", "model-training-artifact", "paper-reference" ]
include
ml-contribution-v5
5
[ "car-racing", "cma-es", "echo-state-networks", "gym", "mdnrnn", "ppo", "proximal-policy-optimization", "rcrc", "reinforcement-learning", "reservoir-computing", "rnn", "vae-pytorch", "variational-autoencoder", "world-models" ]
2025-05-28T04:16:27Z
https://github.com/Shahdsaf/Semi-Supervised-World-Models
null
[ "deep-learning", "natural-language-processing", "representation-learning" ]
[ "contrastive-learning", "self-supervised-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "general.contrastive-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-11-26T18:36:43Z
The corresponding code from our paper "DeCLUTR: Deep Contrastive Learning for Unsupervised Textual Representations". Do not hesitate to open an issue if you run into any trouble!
[ "deep-learning", "natural-language-processing", "representation-learning" ]
[ "representation-learning", "self-supervised-learning", "embedding", "transformer" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
33
224,261,160
https://aclanthology.org/2021.acl-long.72/
Python
Apache-2.0
[ "contrastive-learning", "self-supervised-learning" ]
JohnGiorgi/DeCLUTR
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
2
2026-09-28T18:50:00Z
[]
2023-04-21T01:57:07Z
[ "general.contrastive-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
376
[ "allennlp", "contrastive-learning", "metric-learning", "natural-language-processing", "pytorch", "representation-learning", "self-supervised-learning", "semantic-search", "semantic-text-similarity", "sentence-embeddings", "sentence-similarity", "transformers" ]
2026-08-20T05:59:37Z
https://github.com/JohnGiorgi/DeCLUTR
null
[ "general-ml" ]
[ "novel-method", "support-vector-machine" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.novel-method" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-11-27T01:18:51Z
SGL-SVM: a novel method for tumor classification via support vector machine with sparse group Lasso
[ "general-ml" ]
[ "classifier" ]
ml_related_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
2
224,317,097
null
R
null
[ "novel-method", "support-vector-machine" ]
QUST-AIBBDRC/SGL-SVM
[ "description", "query-match", "repository-metadata" ]
4
2026-09-25T16:01:25Z
[]
2019-11-27T01:26:51Z
[ "general.novel-method" ]
null
2026-09-26T15:53:33Z
gh-ml-readme-evidence-v2
null
[]
[]
missing
2026-09-26T15:53:33Z
QUST-AIBBDRC/SGL-SVM
specific-method-with-novelty-claim
[ "method-tied-novelty-claim", "ml-method-cue" ]
include
ml-contribution-v5
3
[]
2026-08-16T01:25:53Z
https://github.com/QUST-AIBBDRC/SGL-SVM
null
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[ "neural-network-pruning", "pruning", "structured-pruning" ]
[ "description", "paper-reference", "query-match", "repository-metadata" ]
[ "efficiency.pruning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-12-04T02:19:55Z
Pytorch implementation of our paper accepted by IJCAI 2020 -- Channel Pruning via Automatic Structure Search
[ "efficient-ml", "machine-learning-systems", "model-compression" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
27
225,759,924
https://arxiv.org/abs/2001.08565
Python
null
[ "neural-network-pruning", "pruning", "structured-pruning" ]
lmbxmu/ABCPruner
[ "description", "paper-reference", "query-match", "repository-metadata" ]
4
2026-09-29T17:30:08Z
[]
2021-02-11T16:54:47Z
[ "efficiency.pruning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
147
[]
2026-02-10T08:00:53Z
https://github.com/lmbxmu/ABCPruner
null
[ "efficient-ml", "model-compression" ]
[ "distillation", "knowledge-distillation" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "efficiency.knowledge-distillation" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-12-05T09:14:18Z
[AAAI-2020] Official implementation for "Online Knowledge Distillation with Diverse Peers".
[ "efficient-ml", "model-compression" ]
[ "deep-learning", "machine-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
14
226,061,776
null
Python
null
[ "distillation", "knowledge-distillation" ]
DefangChen/OKDDip
[ "description", "github-topics", "query-match", "repository-metadata" ]
7
2026-09-26T14:50:46Z
[]
2023-07-06T21:27:36Z
[ "efficiency.knowledge-distillation" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue" ]
include
ml-contribution-v5
76
[ "deep-learning", "knowledge-distillation", "machine-learning" ]
2026-01-01T04:42:01Z
https://github.com/DefangChen/OKDDip
null
[ "deep-learning", "graph-learning" ]
[ "graph-classification", "graph-representation-learning", "graph-transformer", "message-passing", "transformer" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "graph.graph-classification", "graph.graph-transformer" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-12-16T11:27:15Z
IEEE TNNLS 2021, transformer, multi-graph transformer, graph, graph classification, sketch recognition, sketch classification, free-hand sketch, official code of the paper "Multi-Graph Transformer for Free-Hand Sketch Recognition"
[ "deep-learning", "graph-learning" ]
[ "classifier", "transformer" ]
ml_related_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
32
228,371,818
null
Python
MIT
[ "graph-classification", "graph-representation-learning", "graph-transformer", "message-passing", "transformer" ]
PengBoXiangShang/multigraph_transformer
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
4
2026-09-25T16:01:25Z
[]
2021-05-10T07:25:14Z
[ "graph.graph-classification", "graph.graph-transformer" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
305
[ "graph", "multi-graph-transformer", "pytorch", "pytorch-implementation", "sketch", "sketch-recognition", "sparse-graphs", "transformer", "transformer-architecture" ]
2026-08-18T13:39:26Z
https://github.com/PengBoXiangShang/multigraph_transformer
null
[ "distributed-ml", "privacy-and-federated-learning" ]
[ "collaborative-learning", "federated-learning" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "trust.federated-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-12-20T05:35:53Z
This is the code for our paper `Robust Federated Learning with Attack-Adaptive Aggregation' accepted by FTL-IJCAI'21.
[ "distributed-ml", "privacy-and-federated-learning" ]
[ "machine-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
8
229,197,109
null
Jupyter Notebook
null
[ "collaborative-learning", "federated-learning" ]
cpwan/Attack-Adaptive-Aggregation-in-Federated-Learning
[ "description", "github-topics", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2023-06-12T21:29:28Z
[ "trust.federated-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
46
[ "adversarial-attacks", "federated-learning", "machine-learning", "robust-machine-learning" ]
2026-09-16T14:48:14Z
https://github.com/cpwan/Attack-Adaptive-Aggregation-in-Federated-Learning
null
[ "deep-learning", "representation-learning" ]
[ "continual-learning", "lifelong-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "general.continual-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2019-12-30T07:03:51Z
Official code for ICLR 2020 paper "A Neural Dirichlet Process Mixture Model for Task-Free Continual Learning."
[ "deep-learning", "representation-learning" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-26T14:50:46Z
false
16
230,865,834
null
Python
MIT
[ "continual-learning", "lifelong-learning" ]
soochan-lee/CN-DPM
[ "description", "license-metadata", "query-match", "repository-metadata" ]
3
2026-09-29T17:30:08Z
[]
2020-08-22T01:05:18Z
[ "general.continual-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-context-only", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
100
[]
2026-09-07T07:58:02Z
https://github.com/soochan-lee/CN-DPM
null
[ "biology", "computational-biology", "microscopy", "robotics-and-control" ]
[ "cell-segmentation", "computer-vision", "image-analysis" ]
[ "description", "query-match", "repository-metadata" ]
[ "bio.microscopy-cell-analysis" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-01-06T17:30:48Z
Python code for recurrent fully convolutional network (RFCN) models from our paper "Deep learning robotic guidance for autonomous vascular access"
[ "biology", "computational-biology", "microscopy", "robotics-and-control" ]
[ "deep-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T14:50:46Z
false
0
232,151,079
null
Python
null
[ "cell-segmentation", "computer-vision", "image-analysis" ]
alvchn/nmi-vasc-robot
[ "description", "query-match", "repository-metadata" ]
1
2026-09-26T14:50:46Z
[]
2025-02-04T03:31:12Z
[ "bio.microscopy-cell-analysis" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
1
[]
2025-02-04T17:25:34Z
https://github.com/alvchn/nmi-vasc-robot
null
[ "general-ml" ]
[ "novel-method" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.novel-method" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-01-07T19:40:31Z
Implementation of a novel method for accelerating molecular dynamics with support vector machines I developed in Python using scikit-learn. This repo also includes the results for this project.
[ "general-ml" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-28T18:50:00Z
false
0
232,402,842
null
null
null
[ "novel-method" ]
jrduncan831/Accelerating-Molecular-Dynamics-with-Support-Vector-Machines
[ "description", "query-match", "repository-metadata" ]
1
2026-09-28T18:50:00Z
[]
2020-01-08T17:17:19Z
[ "general.novel-method" ]
d782ebf3e463817ad1afc218ebe3c9ae0f7ed106
2026-09-28T20:05:39Z
gh-ml-readme-evidence-v2
"d782ebf3e463817ad1afc218ebe3c9ae0f7ed106"
[ "other", "overview" ]
[ "paper-reference" ]
ok
2026-09-28T20:05:39Z
jrduncan831/Accelerating-Molecular-Dynamics-with-Support-Vector-Machines
specific-method-with-novelty-claim
[ "contribution-language", "method-tied-novelty-claim", "ml-method-cue", "paper-reference" ]
include
ml-contribution-v5
1
[]
2025-08-04T01:16:30Z
https://github.com/jrduncan831/Accelerating-Molecular-Dynamics-with-Support-Vector-Machines
null
[ "deep-learning", "graph-learning" ]
[ "graph-neural-network", "message-passing", "neural-network" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "graph.gnn-description" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-01-13T18:10:26Z
Code for Paper Graph Neural Networks for Image Understanding Based on Multiple Cues: Group Emotion Recognition and Event Recognition as Use Cases
[ "deep-learning", "graph-learning" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-29T17:30:08Z
false
4
233,661,333
null
Python
MIT
[ "graph-neural-network", "message-passing", "neural-network" ]
gxstudy/Graph-Neural-Networks-for-Image-Understanding-Based-on-Multiple-Cues
[ "description", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-29T17:30:08Z
[]
2020-06-27T20:13:51Z
[ "graph.gnn-description" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
15
[]
2025-11-03T12:42:45Z
https://github.com/gxstudy/Graph-Neural-Networks-for-Image-Understanding-Based-on-Multiple-Cues
null
[ "deep-learning", "representation-learning" ]
[ "distillation", "knowledge-distillation", "representation-learning", "self-supervised-learning" ]
[ "description", "github-topics", "query-match", "repository-metadata" ]
[ "general.self-supervised-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-02-02T10:23:51Z
The source code of our IJCAI 2018 paper "Better and Faster: Knowledge Transfer from Multiple Self-supervised Learning Tasks via Graph Distillation for Video Classification".
[ "deep-learning", "representation-learning" ]
[ "self-supervised-learning", "classifier" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
1
237,753,556
null
null
null
[ "distillation", "knowledge-distillation", "representation-learning", "self-supervised-learning" ]
zcrwind/ss-graph-distillation
[ "description", "github-topics", "query-match", "repository-metadata" ]
2
2026-09-28T18:50:00Z
[]
2020-02-02T10:23:52Z
[ "general.self-supervised-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
1
[ "graph", "ijcai-18", "knowledge-distillation", "pytorch", "self-supervised-learning" ]
2021-09-29T18:44:34Z
https://github.com/zcrwind/ss-graph-distillation
null
[ "deep-learning", "representation-learning" ]
[ "representation-learning", "self-supervised-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "general.self-supervised-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-02-03T03:47:53Z
self-supervised learning, deep learning, representation learning, RotNet, temporal convolutional network(TCN), deformation transformation, sketch pre-train, sketch classification, sketch retrieval, free-hand sketch, official code of paper "Deep Self-Supervised Representation Learning for Free-Hand Sketch"
[ "deep-learning", "representation-learning" ]
[ "deep-learning", "representation-learning", "self-supervised-learning", "classifier" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T20:57:28.489717Z
false
13
237,879,732
null
Python
MIT
[ "representation-learning", "self-supervised-learning" ]
zzz1515151/self-supervised_learning_sketch
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
2
2026-09-27T16:09:59Z
[]
2020-02-24T01:29:56Z
[ "general.self-supervised-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "contribution-language", "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
48
[ "deep-learning", "free-hand-sketch", "representation-learning", "rotnet", "self-supervised", "self-supervised-learning", "sketch-classificaton", "sketch-recognition", "sketch-retrieval", "temporal-convolutional-network", "temporal-convolutions" ]
2025-05-27T18:12:18Z
https://github.com/zzz1515151/self-supervised_learning_sketch
null
[ "classification", "deep-learning", "representation-learning", "time-series", "time-series-and-forecasting" ]
[ "classification", "representation-learning", "self-supervised-learning", "sequence-modeling" ]
[ "description", "query-match", "repository-metadata" ]
[ "general.self-supervised-learning", "timeseries.classification" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-02-13T16:07:27Z
Code for our paper Self Supervised Learning for Semi Supervised Time Series Classification PAKDD 2020
[ "deep-learning", "representation-learning", "time-series-and-forecasting" ]
[ "self-supervised-learning", "classifier" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
7
240,302,639
null
Python
null
[ "representation-learning", "self-supervised-learning" ]
super-shayan/semi-super-ts-clf
[ "description", "query-match", "repository-metadata" ]
4
2026-09-27T16:09:59Z
[]
2020-09-21T18:25:18Z
[ "general.self-supervised-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
16
[]
2024-06-25T11:45:14Z
https://github.com/super-shayan/semi-super-ts-clf
null
[ "reinforcement-learning", "robotics" ]
[ "domain-randomization", "reinforcement-learning", "sim-to-real" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "robotics.domain-randomization" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-02-14T10:48:54Z
Code associated with our paper "Robust Visual Domain Randomization for Reinforcement Learning"
[ "reinforcement-learning", "robotics" ]
[ "reinforcement-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
1
240,486,813
null
Python
MIT
[ "domain-randomization", "reinforcement-learning", "sim-to-real" ]
IndustAI/visual-domain-randomization
[ "description", "license-metadata", "query-match", "repository-metadata" ]
3
2026-09-25T16:01:25Z
[]
2022-11-22T04:39:24Z
[ "robotics.domain-randomization" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
5
[]
2021-03-08T08:59:04Z
https://github.com/IndustAI/visual-domain-randomization
null
[ "forecasting", "time-series", "time-series-and-forecasting" ]
[ "forecasting", "sequence-modeling" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "timeseries.forecasting" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-02-25T02:43:54Z
Code for paper: Block Hankel Tensor ARIMA for Multiple Short Time Series Forecasting (AAAI-20)
[ "forecasting", "time-series", "time-series-and-forecasting" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
40
242,893,533
null
Python
MIT
[ "forecasting", "sequence-modeling" ]
huawei-noah/BHT-ARIMA
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2021-06-08T01:29:04Z
[ "timeseries.forecasting" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
106
[ "arima-forecasting", "tensor-decomposition", "tensor-factorization", "time-series" ]
2026-09-07T07:58:07Z
https://github.com/huawei-noah/BHT-ARIMA
null
[ "interpretability-and-safety", "machine-learning-systems", "trustworthy-ml" ]
[ "adversarial-robustness", "neural-network", "quantization", "robust-optimization" ]
[ "description", "query-match", "repository-metadata" ]
[ "trust.adversarial-robustness" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-03-11T09:25:08Z
Code for paper: "Impact of Low-bitwidth Quantization on the Adversarial Robustness for Embedded Neural Networks"
[ "interpretability-and-safety", "machine-learning-systems", "trustworthy-ml" ]
[ "neural-network" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-27T16:09:59Z
false
0
246,527,624
null
Python
null
[ "adversarial-robustness", "neural-network", "quantization", "robust-optimization" ]
RemiBERNHARD/Quantization_Adversarial_Robustness
[ "description", "query-match", "repository-metadata" ]
1
2026-09-27T16:09:59Z
[]
2022-08-19T11:54:25Z
[ "trust.adversarial-robustness" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
3
[]
2025-10-08T12:36:26Z
https://github.com/RemiBERNHARD/Quantization_Adversarial_Robustness
null
[ "automl", "computer-vision", "general-ml" ]
[ "few-shot-learning", "meta-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "general.meta-learning" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-03-31T04:09:48Z
Source code for CVPR 2020 paper "Scene-Adaptive Video Frame Interpolation via Meta-Learning"
[ "automl", "computer-vision", "general-ml" ]
[ "deep-learning" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-24T14:28:44Z
false
12
251,497,675
null
Python
MIT
[ "few-shot-learning", "meta learning", "meta-learning" ]
myungsub/meta-interpolation
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-24T14:28:44Z
[]
2020-08-14T07:54:30Z
[ "general.meta-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
80
[ "computer-vision", "cvpr2020", "deep-learning", "frame-interpolation", "meta-learning", "pytorch", "slow-motion", "video-frame-interpolation" ]
2025-03-01T03:57:23Z
https://github.com/myungsub/meta-interpolation
null
[ "classical-ml", "computer-vision", "information-retrieval", "representation-learning" ]
[ "image-retrieval", "metric-learning", "representation-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "general.metric-learning", "vision.image-retrieval" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-04-07T15:09:10Z
Official PyTorch Implementation of Proxy Anchor Loss for Deep Metric Learning, CVPR 2020
[ "classical-ml", "representation-learning" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T14:28:44Z
false
63
253,829,710
null
Python
MIT
[ "metric-learning", "representation-learning" ]
sung-yeon-kim/Proxy-Anchor-CVPR2020
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
5
2026-09-28T18:50:00Z
[]
2022-05-17T01:45:29Z
[ "general.metric-learning" ]
7816b341d5c42f1ce99be28f29551e51c43d6b07
2026-09-25T22:08:11Z
gh-ml-readme-evidence-v1
"7816b341d5c42f1ce99be28f29551e51c43d6b07"
[ "citation", "installation", "method", "other" ]
[ "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
ok
2026-09-25T22:08:11Z
sung-yeon-kim/Proxy-Anchor-CVPR2020
readme-supported-paper-method-implementation
[ "contribution-language", "method-contribution", "ml-method-context", "paper-code-relationship", "paper-reference" ]
include
ml-contribution-v5
326
[ "cvpr2020", "deep-metric-learning", "image-retrieval", "proxy-anchor-loss", "pytorch" ]
2026-09-23T18:01:13Z
https://github.com/sung-yeon-kim/Proxy-Anchor-CVPR2020
null
[ "information-retrieval", "software-engineering" ]
[ "code-retrieval", "convolutional-neural-network", "neural-network", "representation-learning" ]
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
[ "growth26.software-neural-code-search" ]
false
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-04-10T17:35:23Z
This repository contains the implementations of our experiments and our approach presented in the paper: CoNCRA: A Convolutional Neural Network Code Retrieval Approach
[ "information-retrieval", "software-engineering" ]
[ "neural-network", "embedding" ]
direct_ml_text
gh-ml-relevance-v1
2026-09-26T22:51:24Z
false
3
254,696,488
null
Jupyter Notebook
MIT
[ "code-retrieval", "convolutional-neural-network", "neural-network", "representation-learning" ]
mrezende/concra
[ "description", "github-topics", "license-metadata", "query-match", "repository-metadata" ]
1
2026-09-26T22:51:24Z
[]
2021-02-12T16:43:16Z
[ "growth26.software-neural-code-search" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-context-only", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
5
[ "code-retrieval", "code-search", "joint-embedding", "neural-networks" ]
2024-02-21T09:58:18Z
https://github.com/mrezende/concra
null
[ "distributed-ml", "privacy-and-federated-learning" ]
[ "collaborative-learning", "federated-learning" ]
[ "description", "license-metadata", "query-match", "repository-metadata" ]
[ "trust.federated-learning" ]
true
candidate
ml-candidate-v3
true
selected-by-current-rule
2020-04-11T23:12:32Z
Code for paper "Adaptive Federated Learning in Resource Constrained Edge Computing Systems"
[ "distributed-ml", "privacy-and-federated-learning" ]
[]
no_text_signal
gh-ml-relevance-v1
2026-09-24T10:49:00Z
false
83
254,971,503
null
Python
MIT
[ "collaborative-learning", "federated-learning" ]
IBM/adaptive-federated-learning
[ "description", "license-metadata", "query-match", "repository-metadata" ]
5
2026-09-28T18:50:00Z
[]
2025-05-07T23:59:52Z
[ "trust.federated-learning" ]
null
null
null
null
null
null
null
null
null
official-paper-method-implementation
[ "ml-method-cue", "official-paper-implementation-cue", "paper-and-code-cue" ]
include
ml-contribution-v5
292
[]
2026-09-25T08:12:29Z
https://github.com/IBM/adaptive-federated-learning
null
End of preview. Expand in Data Studio

GitHub ML

A continually refreshed registry of GitHub repositories across ML fields. The curated current view applies ml-contribution-v5 and seeks projects that present a distinct contribution to an ML model, method, or technique. Broad Search, topic, census, and paper-link discovery is retained as raw provenance; Search observations take precedence during current-view selection.

The selector is a high-precision text heuristic, not verification. Self-description cannot establish actual novelty, correctness, reproducibility, or scientific quality. Its rules screen common forks, owner/profile repositories, coursework, resource collections, tutorial-only projects, and unrelated utilities from the curated views. Qualified review rows are available in the named candidates configuration; reproductions, applications, utilities, and paper associations may provide review context, but none alone implies a distinct contribution or novelty. The complete append-only retrieval history remains available in the opt-in observations configuration for audit. Metadata can be incomplete or stale, and the registry is not a comprehensive census of ML work.

The display name is GitHub ML; gh-ml is the dataset and code repository slug.

Data files

The collector appends machine-readable JSON Lines observations by UTC publication date, with run coverage and resumable checkpoints, using this layout:

README.md                            # uploaded with the first current snapshot
data/current/repositories.parquet  # strict current view
data/candidates/repositories.parquet # included plus plausible review candidates
data/repositories/repositories.parquet # selected current-view row per non-fork GitHub ID
data/history/observations.parquet  # Parquet projection for the observations config
data/current/manifest.json         # v9 source and view counts, hashes, and provenance
data/observations/YYYY/MM/DD/<run-id>.jsonl
data/readme-evidence/YYYY/MM/DD/<run-id>.jsonl # compact README evidence only
data/readme-evidence/YYYY/MM/DD/<run-id>.coverage.json
data/readme-evidence/YYYY/MM/DD/<run-id>.manifest.json
coverage/<run-id>.json
state/checkpoint.json
state/sample.json
state/historical-sample.json
state/backfill.json
state/backfill-fair.json
state/readme-evidence.json
state/census.json                    # queryless Core API cursor and durable state
state/topic-breadth.json             # topic GraphQL cursors and 30-day sweep state
coverage/topic-breadth-<run-id>.json
runs/topic-breadth-<run-id>.manifest.json
coverage/census-<run-id>.json        # census page coverage for each run
runs/census-<run-id>.manifest.json   # published census run marker

Each JSONL row is one raw repository observation, not a unique repository across the full history. Daily Search passes append observations and coverage. The queryless census-daily pass independently walks bounded pages from Core /repositories; it appends candidate-only enriched observations and page coverage, and stores its cursor and resumable state on the Hub. The configured daily census maximum is 50 pages. GitHub Search excludes forks by default; query qualifiers are preserved when explicitly configured. Snapshot publication derives data/history/observations.parquet from append-only JSONL without applying the novelty selector, so the observations config remains an unfiltered audit history. For each numeric github_id, Search observations take precedence over topic and census observations; within the selected source, the greatest observed_at is used. The projection then applies ml-contribution-v5: data/current/repositories.parquet contains include rows, while data/candidates/repositories.parquet contains includes plus qualified review rows. The opt-in repositories config contains one selected current-view row per numeric GitHub ID whose row has fork: false (Search takes precedence over topic and census; latest within the selected source); it does not apply the novelty selector, establish a verified novel ML set, or group full fork families. current seeks projects presenting a distinct ML model, method, or technique contribution. A reproduction, application, utility, or paper association can be retained in raw observations or provide context for a qualified candidate review, but does not alone imply novelty. Manifest version 9 records input revisions and hashes, raw observation-row count and Parquet hash, distinct latest-repository count, strict include/review/exclude counts, and the non-fork view count and hash. state/sample.json stores breadth-sample progress, state/historical-sample.json annual-sample progress, state/checkpoint.json recent collection progress, state/backfill-fair.json fair backfill progress, state/readme-evidence.json README refresh progress, and state/census.json the census cursor and run state; legacy state/backfill.json remains available for historical runs. Coverage records attempted query/date partitions and census pages, counts, outcomes, unresolved enrichment, and known gaps. Search, census, and topic observations and coverage are retained as run history, while checkpoints are replaced as collection continues. The configured daily workflow has four bounded Search passes, a bounded README-evidence pass, queryless census and topic passes, a bounded Hugging Face Daily Papers pass, and a snapshot step. See the topic breadth guide. See the census guide, Hugging Face repository structure, and dataset cards.

The Hugging Face Daily Papers pass uses a bounded recent replay and historical backfill, controlled by a dispatch page budget from 1 to 100 (default 20), plus up to 400 individual paper-detail requests by default. The list endpoint omits githubRepo; bounded detail hydration uses a separate queue and checkpoint, with a configurable budget of 0 to 1000 requests per run. Page, detail, and GitHub GraphQL lookup budgets are separate, and coverage records their counts and remaining detail work. Its user-submitted GitHub links are unverified discovery signals, do not establish official implementations, and do not bypass the strict novelty selector for the default current view. Paper-link assertions retain separate provenance under data/paper-links/; any linked repository observations remain in the raw observation history and are selected by the usual projection rules. See the Hugging Face Daily Papers guide for checkpoint behavior and limits.

The README pass selects at most 150 eligible repositories per run through GitHub’s Core API and publishes compact extracted evidence and its separate checkpoint. The daily pass makes at most 150 GitHub Core API README requests; successful 200/304 checks are revisited after 365 days to protect first-touch coverage, 404 responses after 30 days, and transient errors after one day, while repository renames trigger an immediate refetch. The census pass publishes append-only candidate observations, per-run page coverage, and resumable state in its own Hub commit. The snapshot publisher then commits data/current/repositories.parquet, data/candidates/repositories.parquet, data/repositories/repositories.parquet, data/history/observations.parquet, this card, and data/current/manifest.json together in one atomic Hub commit from the observation history and available README evidence. The configured workflow rebuilds all four Parquet files from observations and available README evidence on main; a manual snapshot_only dispatch can refresh them without running collectors. If a collector fails, the snapshot step still derives from successful collector commits, then the workflow reports the failure at its final gate. The default Parquet contains one selected include row per GitHub ID; review and exclude rows are omitted. The repositories Parquet contains one selected current-view row per ID only when fork is exactly false. The observations Parquet retains every raw observation row without selector filtering or README enrichment; available README evidence joins into each derived view, including repositories. Manifest version 9 records source files, hashes, counts, selector version, README evidence inputs and fingerprint, and nonfork_repository_count, repositories_parquet_sha256, and repositories_parquet_row_count. Projection version 7 aggregates sorted paper_ids across raw observations and includes the latest compact README evidence available for each repository before evaluating selector v5.

Fields

Field Meaning
github_id Stable numeric GitHub repository ID and deduplication key
name, url Current repository name and URL
description Repository description, nullable
created_at, updated_at, pushed_at GitHub timestamps
stars, forks Observed repository counts
language, license, topics, homepage GitHub metadata; nullable or empty when missing
archived, fork Repository state flags; GitHub Search excludes forks by default
domains, methods Multi-label tags represented as lists of lowercase hyphenated slugs
query_ids Collection queries that matched the repository
paper_ids Sorted union of associated Hugging Face Daily Papers IDs across raw observations; link association is unverified and does not prove an official implementation or novelty
observed_at UTC timestamp for this metadata snapshot
novelty_signals Evidence labels such as query match, paper reference, or model weights; not novelty verification
candidate_status Queryless census/topic discovery status (candidate, unknown, or not_candidate); not novelty verification
candidate_rule_version Version of the rule that marks rows eligible for the candidates view; currently ml-candidate-v3
candidate_eligible Whether the row is in the candidates view; strict includes and qualified review rows are eligible
candidate_reason Reason for the eligibility decision
evidence_version, evidence_tier, evidence_signals Versioned text hints from repository name, description, and topics; they do not verify ML use or novelty.
readme_status, readme_checked_at, readme_evidence_version, readme_signals, readme_sections, readme_blob_sha, readme_etag, readme_repository_name_at_fetch, readme_observed_at Latest compact README evidence and fetch metadata attached by GitHub ID. README text itself is never persisted. Signals support the selector heuristic; they do not verify claims.
selection_version, selection_status, selection_reason, selection_signals Derived current-view fields. The current Parquet contains only include rows; the candidates and repositories Parquets retain their corresponding selection statuses. The manifest reports aggregate review and exclude counts. Raw observations do not contain these decision fields.
first_observed_at, observation_count, all_query_ids, all_domains, all_methods, all_novelty_signals Current-view additions. Counts refer to raw observation history per repository; all_* values are sorted unions across that history.

Labels are open vocabulary, multi-label, and subject to change. They can describe both a field (for example, computer-vision or bioinformatics) and a method (for example, quantization, retrieval, or reinforcement-learning). Missing labels do not mean a project is irrelevant.

Updates and deduplication

The configured workflow runs four Search passes with a default total budget of 2,000 Search requests, followed by up to 150 README fetch requests through GitHub Core API: breadth sample (500), recent collection (600), annual historical sample (200), and fair historical backfill (700), subject to Actions timeout and GitHub API limits. The Search catalog has 584 queries (569 existing plus 15 additions) and continues to evolve. Its broad matches form raw provenance; GitHub Search excludes forks by default, and explicit query qualifiers are preserved. The breadth pass rotates through the catalog with one created: first-page search per selected query, so a daily budget may leave some queries untouched and ranking can omit matches. The recent pass spends the same 600-request budget round-robin across query lanes; each Search request advances that lane’s pushed-date cursor, which resumes independently per query. Search caps, incomplete results, indexing gaps, and the bounded budget still limit coverage. The README evidence pass is a separate, bounded Core API enrichment and does not add Search queries or change query provenance. These retrieval passes improve recall in the audit history but do not determine the default dataset or verify novelty.

The historical-sample pass takes one ranked first-page search for every query/year lane from 2008 through a campaign end date fixed when the campaign starts. That end date stays fixed across partial runs and later runs with an expanded catalog. Its v2 completion ledger preserves completed (query ID, query text, year) lanes across catalog additions and edits. New or changed queries are sampled for every year in the fixed campaign, removed queries are dropped, and a legacy v1 cursor is safely migrated by matching query signatures. The checkpoint is state/historical-sample.json on the Hub and historical-sample-state.json in local output. Completed ledger entries persist, so future catalog additions resume only their missing lanes. The bounded request budget remains 200 per scheduled day, processed round-robin across query groups. These annual samples can improve breadth across creation years, but ranked first-page sampling can omit matching repositories; they do not replace historical backfill or establish completeness. The scheduled backfill-fair pass rotates through query/date-partition work, issuing one Search request per query in each rotation to spread its bounded budget across queries. Its checkpoints are state/backfill-fair.json on the Hub and backfill-fair-state.json in local output. The legacy backfill command still uses state/backfill.json / backfill-state.json; fair backfill uses a separate checkpoint and does not migrate the old cursor. Do not claim exhaustive coverage until all date partitions have been scanned and coverage records show a completed sweep. Recent and breadth checkpoints remain at state/checkpoint.json and state/sample.json, respectively. Within a run, matches merge by numeric github_id, never by mutable name. Across runs, the observation history is append-only; uv run gh-ml current-view /path/to/downloaded-dataset --output ~/.local/share/modelomics-gh-ml/current-view.jsonl builds one row per ID from local data/observations/**/*.jsonl files. It prefers Search observations over topic and census observations, then chooses the greatest observed_at within that source. Any all_* accumulated-label fields retain values across history while ordinary fields come from the chosen row. This local command does not load separate README evidence files; the published snapshot publisher joins available README evidence before selection. The command writes a manifest and does not modify the Hub. For Parquet output, install uv sync --extra parquet and provide --parquet-output <path>.

The current projection prefers Search observations over topic and census observations for each GitHub ID, then selects the greatest observed_at within that source. It adds first_observed_at, observation_count, and sorted unions of labels, including paper_ids, then evaluates repository-owned metadata and any available compact README signals with ml-contribution-v5. Inclusion can come from a method-specific novelty claim with a recognized ML method cue, description text linking an official/authors’ paper and implementation to ML method evidence, or compact README signals that support an ML method contribution with a paper/code relationship or official implementation claim. The candidates view contains every included row plus review rows with ML method and text evidence, a nonempty description, and either a paper-and-code cue or an associated paper_ids value (ml-candidate-v3). A Daily Papers link can therefore make a qualified review row eligible for candidates, but cannot cause inclusion in current; the association is unverified provenance, not proof of an official implementation or novelty. Generic contribution language alone does not qualify. Overview, reproduction, and dataset cues lead to review rather than inclusion on their own; educational cues such as course/coursework, class or course projects, homework, and assignments are screened as coursework, while a bare mention of a class, lecture, or tutorial is not sufficient. Surveys, forks, profiles, and unrelated utilities are screened out, while a utility with a method-specific novelty cue may remain for review. None of these cases alone establishes a distinct contribution or novelty. Only include rows enter current; review rows enter candidates; observations preserves every unfiltered raw observation. The selector uses repository-owned metadata and, when available, extracted README signals, not query labels, to decide. README text is processed in memory and never stored; only bounded enum signals, section names, content hash, ETag, status, and timestamps are published. The README pass prioritizes previously selected include and review rows, applies a 150-request daily maximum, and stores progress in state/readme-evidence.json; compact evidence with an older extraction version is eligible for gradual re-fetch within that same daily budget. Until fetched, old compact signals may remain attached and can still affect selection, so published rows are not all corrected immediately. Evidence refresh is not guaranteed to cover every candidate. README extraction is heuristic and can miss relevant evidence or misread repository claims. The Parquet observation_count counts raw historical observations for each repository, and the manifest reports raw rows, latest repositories, and selection counts. This heuristic cannot verify claims or guarantee novelty; false positives and missed candidates remain possible.

Selector v5 routes exploration or comparison of named existing time-series models to review, even when a repository name says “novel model.” README extraction v2 uses specific educational wording (course/coursework, class or course projects, homework, and assignments); bare mentions of “class,” “lecture,” or “tutorial” do not create a coursework signal. A direct novelty claim with an active README educational signal also stays in review when the active README has that cue without a paper reference or official/paper-code signal; a paper reference is provenance, not proof of officiality or novelty. Stale compact README evidence is re-fetched gradually under the existing 150-request daily cap. Until a repository is fetched, its older compact signal may remain in the published projection.

The source Search catalog currently contains 584 queries (569 existing plus 15 additions). It is intentionally broad and serves only as retrieval provenance for the raw observations config. Query-derived methods or domains cannot satisfy the contribution selector. All records are keyed by numeric github_id. The scheduled census-daily collector independently enumerates bounded pages from GitHub Core /repositories; the README evidence pass remains a distinct Core API enrichment for selected repositories. Search observations take precedence over topic or census observations for IDs seen in those streams, before the strict selector builds the included current and broader candidates views. The census guide and topic breadth guide describe these queryless sources, their state, and limits. The configured topic pass uses GraphQL over an ordered 38-slug catalog, with up to 68 GraphQL pages per daily run: a first-page refresh for each of 38 topics and up to 30 deeper cursor pages. The eight added field topics cover geospatial, remote sensing, bioinformatics, cheminformatics, speech recognition, text to speech, medical imaging, and recommender systems. Completed deep sweeps become eligible to restart after 30 days, while first-page refreshes continue daily; no completeness or snapshot-isolation guarantee is made. Topic collector output and workflow configuration do not establish that a remote run has executed or published.

The source is GitHub's public repository metadata and Search API. GitHub Search caps each query at 1,000 returned results and at 4,000 repositories searched, and is subject to request limits, timeouts, incomplete responses, and indexing gaps. Annual historical sampling covers only the ranked first page for each query/year, so its query/year attempts do not mean it collected every matching repository. A coverage status of capped or incomplete, or coverage_gap: true, flags known gaps; check coverage_gap_reason and the other per-query fields. Broad query coverage and exhaustive backfills improve recall but cannot guarantee exhaustiveness. Results can include false positives, and not all novel ML work is hosted on GitHub or discoverable by the configured queries. See the official Search API documentation.

Access

This dataset is maintained at modelomics/gh-ml. The append-only JSONL observations are the source history on main; the current config provides the strict Parquet view, the candidates config provides the broader discovery view, and observations provides opt-in raw-history Parquet derived from the append-only JSONL source files. Hugging Face Trusted Publisher authentication is configured for repository modelomics/gh-ml, branch main, and workflow daily.yml. The workflow requests id-token: write and exchanges its GitHub identity using HF_OIDC_RESOURCE=datasets/modelomics/gh-ml; see Hugging Face Trusted Publishers. An HF_TOKEN secret, when set, takes precedence over OIDC. Actions supplies GITHUB_TOKEN for GitHub Search. To recover a scheduled Search run, use Run workflow in Actions: each successful pass commits its cursor with observations and coverage, while a failed search leaves the prior checkpoint available for retry.

For local publishing, set HF_TOKEN in the environment or sign in with uv run hf auth login; the collector reads the saved Hugging Face CLI token. GitHub authentication can be provided through GITHUB_TOKEN or gh auth login (the collector reads gh auth token). The HF token must have write permission on modelomics/gh-ml. For local recovery, rerun using the same --output-dir so the local cursor is reused; the default is ~/.local/share/modelomics-gh-ml/runs. Use --no-publish for local-only collection, which writes run files without requiring Hugging Face credentials.

The source card YAML declares four Parquet configs: current is the default strict view, candidates includes plausible review rows, repositories is an opt-in view of selected current-view rows that have fork: false, and observations is opt-in raw history regenerated from the JSONL source files. The repositories view is not a verified novel ML set and does not group full fork families. The snapshot publisher commits all four Parquet artifacts, the manifest, and this card atomically. These configs follow the Hub's dataset repository structure; datasets is needed only by consumers, not by the collector.

from datasets import load_dataset

# Default config: one strict high-precision row per included GitHub ID.
current_default = load_dataset("modelomics/gh-ml")["train"]
current = load_dataset("modelomics/gh-ml", "current")["train"]

# Broader discovery view: strict includes plus plausible review cases.
candidates = load_dataset("modelomics/gh-ml", "candidates")["train"]

# One latest observed row per numeric GitHub ID with fork=false.
repositories = load_dataset("modelomics/gh-ml", "repositories")["train"]

# Opt-in audit history: append-only, unfiltered observations.
observations = load_dataset("modelomics/gh-ml", "observations")["train"]

Coverage is per run; it is not a list of repositories. Breadth sample coverage describes one first page per selected catalog query; annual historical-sample coverage describes one first page per query/year; daily recent coverage describes pushed-date search windows advanced round-robin with per-query cursors; fair and legacy backfill coverage describe created-date partitions. A complete_sweep of false usually means the request budget left a cursor to resume. Within each run's queries, review status, coverage_gap, coverage_gap_reason, incomplete_results, and search_limit_reached. Fair backfill is not evidence of exhaustive coverage until all date partitions have been scanned and coverage records show the completed sweep. Coverage and checkpoint JSON are operational metadata, not rows in the repository table. Hub checkpoints let scheduled or manually dispatched Actions runs resume; locally, reuse the same --output-dir to resume local state. Run the breadth sample locally with uv run gh-ml sample --max-requests 500 --since-days 1; run the annual sample with uv run gh-ml historical-sample --max-requests 200; run fair backfill locally without publishing with uv run gh-ml backfill-fair --max-requests 700 --no-publish. The local fair checkpoint is backfill-fair-state.json and is independent from the legacy backfill checkpoint. Add --no-publish to keep output local.

License and attribution

Repository metadata is sourced from GitHub. Repositories retain their own licenses and terms; this registry does not relicense or redistribute their code. Check the license field and the source repository before reusing any project. Dataset-level licensing should be set to the license selected by the maintainers after review of applicable metadata and policies; license: other above is a placeholder for that decision.

Downloads last month
3,365