Learning from Failure: Training Debiased Classifier from Biased Classifier

Abstract

Neural networks often learn to make predictions that overly rely on spurious correlation existing in the dataset, which causes the model to be biased. While previous work tackles this issue by using explicit labeling on the spuriously correlated attributes or presuming a particular bias type, we instead utilize a cheaper, yet generic form of human knowledge, which can be widely applicable to various types of bias. We first observe that neural networks learn to rely on the spurious correlation only when it is "easier" to learn than the desired knowledge, and such reliance is most prominent during the early phase of training. Based on the observations, we propose a failure-based debiasing scheme by training a pair of neural networks simultaneously. Our main idea is twofold; (a) we intentionally train the first network to be biased by repeatedly amplifying its "prejudice", and (b) we debias the training of the second network by focusing on samples that go against the prejudice of the biased network in (a). Extensive experiments demonstrate that our method significantly improves the training of the network against various types of biases in both synthetic and real-world datasets. Surprisingly, our framework even occasionally outperforms the debiasing methods requiring explicit supervision of the spuriously correlated attributes.

References in corpus (6)

• Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization
• A Closer Look at Memorization in Deep Networks
• Why Can't I Dance in the Mall? Learning to Mitigate Scene Bias in Action Recognition
• Learning De-biased Representations with Biased Representations
• Learning Robust Representations by Projecting Superficial Statistics Out
• Robust Inference via Generative Classifiers for Handling Noisy Labels

Cited by in corpus (21)

• WILDS: A Benchmark of in-the-Wild Distribution Shifts
• Learning Debiased Representation via Disentangled Feature Augmentation
• Gradient Starvation: A Learning Proclivity in Neural Networks
• Fairness via Representation Neutralization
• Can Subnetwork Structure be the Key to Out-of-Distribution Generalization?
• Object-aware Contrastive Learning for Debiased Scene Representation
• Simple data balancing achieves competitive worst-group-accuracy
• Environment Inference for Invariant Learning
• Latent Adversarial Debiasing: Mitigating Collider Bias in Deep Neural Networks
• EnD: Entangling and Disentangling deep representations for bias correction
• Toward Learning Human-aligned Cross-domain Robust Models by Countering Misaligned Features
• BiaSwap: Removing dataset bias with bias-tailored swapping augmentation
• StyleAugment: Learning Texture De-biased Representations by Style Augmentation without Pre-defined Textures
• Learning Stable Classifiers by Transferring Unstable Features
• Unsupervised Learning of Debiased Representations with Pseudo-Attributes
• Encouraging Intra-Class Diversity Through a Reverse Contrastive Loss for Better Single-Source Domain Generalization
• NeuroView: Explainable Deep Network Decision Making
• Fighting Fire with Fire: Contrastive Debiasing without Bias-free Data via Generative Bias-transformation
• Combating Unknown Bias with Effective Bias-Conflicting Scoring and Gradient Alignment
• An Evaluation Dataset and Strategy for Building Robust Multi-turn Response Selection Model
• Unsupervised Learning of Unbiased Visual Representations