Shortcuts - HanjieChen/Reading-List GitHub Wiki

Shortcut Learning

• AGRO: ADVERSARIAL DISCOVERY OF ERROR-PRONE GROUPS FOR ROBUST OPTIMIZATION
• Explainer Divergence Scores (EDS): Some Post-Hoc Explanations May be Effective for Detecting Unknown Spurious Correlations
• Spurious Features Everywhere - Large-Scale Detection of Harmful Spurious Features in ImageNet
• Removing Spurious Features can Hurt Accuracy and Affect Groups Disproportionately
• Shortcut Learning of Large Language Models in Natural Language Understanding: A Survey
• Patch Shortcuts: Interpretable Proxy Models Efficiently Find Black-Box Vulnerabilities
• Informativeness and Invariance: Two Perspectives on Spurious Correlations in Natural Language
• Identifying Spurious Correlations for Robust Text Classification
• Evaluating Models’ Local Decision Boundaries via Contrast Sets
• Combining Feature and Instance Attribution to Detect Artifacts
• Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasets
• Counterfactual Invariance to Spurious Correlations: Why and How to Pass Stress Tests
• Uninformative Input Features and Counterfactual Invariance: Two Perspectives on Spurious Correlations in Natural Language
• FINDING AND FIXING SPURIOUS PATTERNS WITH EXPLANATIONS
• Predictive Biases in Natural Language Processing Models: A Conceptual Framework and Overview
• Towards Debiasing NLU Models from Unknown Biases
• POST HOC EXPLANATIONS MAY BE INEFFECTIVE FOR DETECTING UNKNOWN SPURIOUS CORRELATION
• Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasets
• “Will You Find These Shortcuts?” A Protocol for Evaluating the Faithfulness of Input Salience Methods for Text Classification
• Introspective Distillation for Robust Question Answering
• Towards Domain-Generalizable Paraphrase Identification by Avoiding the Shortcut Learning
• Towards Debiasing Fact Verification Models
• Annotation Artifacts in Natural Language Inference Data
• Hypothesis Only Baselines in Natural Language Inference
• Dataset Cartography: Mapping and Diagnosing Datasets with Training Dynamics
  • [EMNLP 2020] Swabha Swayamdipta, Roy Schwartz, Nicholas Lourie, Yizhong Wang, Hannaneh Hajishirzi, Noah A. Smith, Yejin Choi
• Shortcutted Commonsense: Data Spuriousness in Deep Learning of Commonsense Reasoning
  • [EMNLP 2021] Ruben Branco, António Branco, João António Rodrigues, João Ricardo Silva
• Identifying and Mitigating Spurious Correlations for Improving Robustness in NLP Models
  • [Arxiv 2021 Oct] Tianlu Wang, Diyi Yang, Xuezhi Wang
• Competency Problems: On Finding and Removing Artifacts in Language Data
  • [EMNLP 2021] Matt Gardner, William Merrill, Jesse Dodge, Matthew E. Peters, Alexis Ross, Sameer Singh, Noah A. Smith
• Regularizing Models via Pointwise Mutual Information for Named Entity Recognition
  • [Arxiv 2021 Apr] Minbyul Jeong, Jaewoo Kang
• Issues with Entailment-based Zero-shot Text Classification
  • [ACL 2021] Tingting Ma, Jin-Ge Yao, Chin-Yew Lin, Tiejun Zhao
• Avoiding Inference Heuristics in Few-shot Prompt-based Finetuning
  • [EMNLP 2021] Prasetya Ajie Utama, Nafise Sadat Moosavi, Victor Sanh, Iryna Gurevych
• A Case Study of the Shortcut Effects in Visual Commonsense Reasoning
  • [AAAI 2021] Keren Ye, Adriana Kovashka
• Improving robustness against common corruptions by covariate shift adaptation
  • [NeurIPS 2020] Steffen Schneider, Evgenia Rusak, Luisa Eck, Oliver Bringmann, Wieland Brendel, Matthias Bethge
• Beyond accuracy: quantifying trial-by-trial behaviour of CNNs and humans by measuring error consistency
  • [NeurIPS 2020] Robert Geirhos, Kristof Meding, Felix A. Wichmann
• Unsupervised Learning of Debiased Representations with Pseudo-Attributes
  • [Arxiv 2021 Aug] Seonguk Seo, Joon-Young Lee, Bohyung Han
• Fairness via Representation Neutralization
  • [NeurIPS 2021] Mengnan Du, Subhabrata Mukherjee, Guanchu Wang, Ruixiang Tang, Ahmed Hassan Awadallah, Xia Hu
• Latent Adversarial Debiasing: Mitigating Collider Bias in Deep Neural Networks
  • [Arxiv 2020 Nov] Luke Darlow, Stanisław Jastrzębski, Amos Storkey
• On the Importance of Regularisation & Auxiliary Information in OOD Detection
  • [Arxiv 2021 Jul] John Mitros, Brian Mac Namee
• Debiasing Methods in Natural Language Understanding Make Bias More Accessible
  • [EMNLP 2021] Michael Mendelson, Yonatan Belinkov
• Causally-motivated Shortcut Removal Using Auxiliary Labels
  • [Arxiv 2021 May] Maggie Makar, Ben Packer, Dan Moldovan, Davis Blalock, Yoni Halpern, Alexander D'Amour
• Patch Shortcuts: Interpretable Proxy Models Efficiently Find Black-Box Vulnerabilities
  • [CVPR 2021] Julia Rosenzweig, Joachim Sicking, Sebastian Houben, Michael Mock, Maram Akila
• Why Machine Reading Comprehension Models Learn Shortcuts?
  • [ACL 2021] Yuxuan Lai, Chen Zhang, Yansong Feng, Quzhe Huang, Dongyan Zhao
• Towards Interpreting and Mitigating Shortcut Learning Behavior of NLU Models
  • [NAACL 2021] Mengnan Du, Varun Manjunatha, Rajiv Jain, Ruchi Deshpande, Franck Dernoncourt, Jiuxiang Gu, Tong Sun, Xia Hu
• Shortcut Learning in Deep Neural Networks
  • [Nature 2020 Nov] Robert Geirhos, Jörn-Henrik Jacobsen, Claudio Michaelis, Richard Zemel, Wieland Brendel, Matthias Bethge, Felix A. Wichmann
• Debugging Tests for Model Explanations
  • [NeurIPS 2020] Julius Adebayo, Michael Muelly, Ilaria Liccardi, Been Kim
• Learning Which Features Matter: RoBERTa Acquires a Preference for Linguistic Generalizations (Eventually)
  • [EMNLP 2020] Alex Warstadt, Yian Zhang, Haau-Sing Li, Haokun Liu, Samuel R. Bowman
• The Dangers of Underclaiming: Reasons for Caution When Reporting How NLP Systems Fail