# Learning from others' mistakes: Avoiding dataset biases without modeling them

arXiv: https://arxiv.org/abs/2012.01300 · HTML: https://mlsanity.com/paper/2012.01300/

## Key facts

- Cited by 83 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Dec 2020.

## Built on by (most cited first, 30 of 83)

- [Process for Adapting Language Models to Society (PALMS) with Values-Targeted Datasets](https://mlsanity.com/paper/2106.10328/) (2021)
- [Shortcut Learning of Large Language Models in Natural Language Understanding](https://mlsanity.com/paper/2208.11857/) (2022)
- [Evading the Simplicity Bias: Training a Diverse Set of Models Discovers Solutions with Superior OOD Generalization](https://mlsanity.com/paper/2105.05612/) (2021)
- [Distilling Model Failures as Directions in Latent Space](https://mlsanity.com/paper/2206.14754/) (2022)
- [Generalization in NLI: Ways (Not) To Go Beyond Simple Heuristics](https://mlsanity.com/paper/2110.01518/) (2021)
- [ZIN: When and How to Learn Invariance Without Environment Partition?](https://mlsanity.com/paper/2203.05818/) (2022)
- [Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasets](https://mlsanity.com/paper/2203.12942/) (2022)
- [Supervising Model Attention with Human Explanations for Robust Natural Language Inference](https://mlsanity.com/paper/2104.08142/) (2021)
- [SelecMix: Debiased Learning by Contradicting-pair Sampling](https://mlsanity.com/paper/2211.02291/) (2022)
- [From Shortcuts to Triggers: Backdoor Defense with Denoised PoE](https://mlsanity.com/paper/2305.14910/) (2023)
- [Explanation-based Finetuning Makes Models More Robust to Spurious Cues](https://mlsanity.com/paper/2305.04990/) (2023)
- [End-to-End Self-Debiasing Framework for Robust NLU Training](https://mlsanity.com/paper/2109.02071/) (2021)
- [Avoiding Inference Heuristics in Few-shot Prompt-based Finetuning](https://mlsanity.com/paper/2109.04144/) (2021)
- [BLIND: Bias Removal With No Demographics](https://mlsanity.com/paper/2212.10563/) (2022)
- [Causal Balancing for Domain Generalization](https://mlsanity.com/paper/2206.05263/) (2022)
- [Context-aware Adversarial Training for Name Regularity Bias in Named Entity Recognition](https://mlsanity.com/paper/2107.11610/) (2021)
- [RAIL-KD: RAndom Intermediate Layer Mapping for Knowledge Distillation](https://mlsanity.com/paper/2109.10164/) (2021)
- [Debiasing Methods in Natural Language Understanding Make Bias More Accessible](https://mlsanity.com/paper/2109.04095/) (2021)
- [Predict then Interpolate: A Simple Algorithm to Learn Stable Classifiers](https://mlsanity.com/paper/2105.12628/) (2021)
- [Uncertainty Calibration for Ensemble-Based Debiasing Methods](https://mlsanity.com/paper/2111.04104/) (2021)
- [Learning to Split for Automatic Bias Detection](https://mlsanity.com/paper/2204.13749/) (2022)
- [Debiasing Masks: A New Framework for Shortcut Mitigation in NLU](https://mlsanity.com/paper/2210.16079/) (2022)
- [Causal-Guided Active Learning for Debiasing Large Language Models](https://mlsanity.com/paper/2408.12942/) (2024)
- [Are Bias Mitigation Techniques for Deep Learning Effective?](https://mlsanity.com/paper/2104.00170/) (2021)
- [Robustness Challenges in Model Distillation and Pruning for Natural Language Understanding](https://mlsanity.com/paper/2110.08419/) (2021)
- [OccamNets: Mitigating Dataset Bias by Favoring Simpler Hypotheses](https://mlsanity.com/paper/2204.02426/) (2022)
- [Kernel-Whitening: Overcome Dataset Bias with Isotropic Sentence Embedding](https://mlsanity.com/paper/2210.07547/) (2022)
- [Beyond Distribution Shift: Spurious Features Through the Lens of Training Dynamics](https://mlsanity.com/paper/2302.09344/) (2023)
- [Analyzing the Evaluation of Cross-Lingual Knowledge Transfer in Multilingual Language Models](https://mlsanity.com/paper/2402.02099/) (2024)
- [AGRO: Adversarial Discovery of Error-prone groups for Robust Optimization](https://mlsanity.com/paper/2212.00921/) (2022)

## Abstract (arXiv)

State-of-the-art natural language processing (NLP) models often learn to model dataset biases and surface form correlations instead of features that target the intended underlying task. Previous work has demonstrated effective methods to circumvent these issues when knowledge of the bias is available. We consider cases where the bias issues may not be explicitly identified, and show a method for training models that learn to ignore these problematic correlations. Our approach relies on the observation that models with limited capacity primarily learn to exploit biases in the dataset. We can leverage the errors of such limited capacity models to train a more robust model in a product of experts, thus bypassing the need to hand-craft a biased model. We show the effectiveness of this method to retain improvements in out-of-distribution settings even if no particular bias is targeted by the biased model.