Learning Debiased Representation via Disentangled Feature Augmentation

Abstract

Image classification models tend to make decisions based on peripheral attributes of data items that have strong correlation with a target variable (i.e., dataset bias). These biased models suffer from the poor generalization capability when evaluated on unbiased datasets. Existing approaches for debiasing often identify and emphasize those samples with no such correlation (i.e., bias-conflicting) without defining the bias type in advance. However, such bias-conflicting samples are significantly scarce in biased datasets, limiting the debiasing capability of these approaches. This paper first presents an empirical analysis revealing that training with "diverse" bias-conflicting samples beyond a given training set is crucial for debiasing as well as the generalization capability. Based on this observation, we propose a novel feature-level data augmentation technique in order to synthesize diverse bias-conflicting samples. To this end, our method learns the disentangled representation of (1) the intrinsic attributes (i.e., those inherently defining a certain class) and (2) bias attributes (i.e., peripheral attributes causing the bias), from a large number of bias-aligned samples, the bias attributes of which have strong correlation with the target variable. Using the disentangled representation, we synthesize bias-conflicting samples that contain the diverse intrinsic attributes of bias-aligned samples by swapping their latent features. By utilizing these diversified bias-conflicting features during the training, our approach achieves superior classification accuracy and debiasing results against the existing baselines on synthetic and real-world datasets.

arXiv commentAccepted to NeurIPS 2021 as Oral Presentation

References in corpus (11)

• Shortcut Learning in Deep Neural Networks
• ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness
• Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization
• Approximating CNNs with Bag-of-local-Features models works surprisingly well on ImageNet
• Learning De-biased Representations with Biased Representations
• Learning Robust Representations by Projecting Superficial Statistics Out
• Learning from Failure: Training Debiased Classifier from Biased Classifier
• FairFil: Contrastive Neural Debiasing Method for Pretrained Text Encoders
• Self-Challenging Improves Cross-Domain Generalization
• Domain Generalization with MixStyle
• Latent Adversarial Debiasing: Mitigating Collider Bias in Deep Neural Networks

Cited by in corpus (5)

• Predicting Scientific Impact Through Diffusion, Conformity, and Contribution Disentanglement
• Echoes: Unsupervised Debiasing via Pseudo-bias Labeling in an Echo Chamber
• Unsupervised Learning of Unbiased Visual Representations
• 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