Unlearn Dataset Bias in Natural Language Inference by Fitting the Residual

Abstract

Statistical natural language inference (NLI) models are susceptible to learning dataset bias: superficial cues that happen to associate with the label on a particular dataset, but are not useful in general, e.g., negation words indicate contradiction. As exposed by several recent challenge datasets, these models perform poorly when such association is absent, e.g., predicting that "I love dogs" contradicts "I don't love cats". Our goal is to design learning algorithms that guard against known dataset bias. We formalize the concept of dataset bias under the framework of distribution shift and present a simple debiasing algorithm based on residual fitting, which we call DRiFt. We first learn a biased model that only uses features that are known to relate to dataset bias. Then, we train a debiased model that fits to the residual of the biased model, focusing on examples that cannot be predicted well by biased features only. We use DRiFt to train three high-performing NLI models on two benchmark datasets, SNLI and MNLI. Our debiased models achieve significant gains over baseline models on two challenge test sets, while maintaining reasonable performance on the original test sets.

arXiv commentDeepLo at EMNLP 2019

References in corpus (10)

• Adam: A Method for Stochastic Optimization
• Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
• Enhanced LSTM for Natural Language Inference
• Theoretically Principled Trade-off between Robustness and Accuracy
• Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them
• On Causal and Anticausal Learning
• Does Distributionally Robust Supervised Learning Give Robust Classifiers?
• Learning Robust Representations by Projecting Superficial Statistics Out
• Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference
• WinoGrande: An Adversarial Winograd Schema Challenge at Scale

Cited by in corpus (2)

• Don't Take the Easy Way Out: Ensemble Based Methods for Avoiding Known Dataset Biases
• Evaluating Gender Bias in Natural Language Inference