Fairness via Representation Neutralization

Abstract

Existing bias mitigation methods for DNN models primarily work on learning debiased encoders. This process not only requires a lot of instance-level annotations for sensitive attributes, it also does not guarantee that all fairness sensitive information has been removed from the encoder. To address these limitations, we explore the following research question: Can we reduce the discrimination of DNN models by only debiasing the classification head, even with biased representations as inputs? To this end, we propose a new mitigation technique, namely, Representation Neutralization for Fairness (RNF) that achieves fairness by debiasing only the task-specific classification head of DNN models. To this end, we leverage samples with the same ground-truth label but different sensitive attributes, and use their neutralized representations to train the classification head of the DNN model. The key idea of RNF is to discourage the classification head from capturing spurious correlation between fairness sensitive information in encoder representations with specific class labels. To address low-resource settings with no access to sensitive attribute annotations, we leverage a bias-amplified model to generate proxy annotations for sensitive attributes. Experimental results over several benchmark datasets demonstrate our RNF framework to effectively reduce discrimination of DNN models with minimal degradation in task-specific performance.

arXiv commentAccepted by NeurIPS 2021

References in corpus (14)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Distilling the Knowledge in a Neural Network
• Shortcut Learning in Deep Neural Networks
• Counterfactual Fairness
• Manifold Mixup: Better Representations by Interpolating Hidden States
• Invariant Risk Minimization
• AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias
• Achieving Fairness through Adversarial Learning: an Application to Recidivism Prediction
• Learning from Failure: Training Debiased Classifier from Biased Classifier
• FairFil: Contrastive Neural Debiasing Method for Pretrained Text Encoders
• Gradient Starvation: A Learning Proclivity in Neural Networks
• Probing Neural Network Comprehension of Natural Language Arguments
• Repairing without Retraining: Avoiding Disparate Impact with Counterfactual Distributions
• Fair Mixup: Fairness via Interpolation

Cited by in corpus (6)

• Fair Oversampling Technique using Heterogeneous Clusters
• Toward Operationalizing Pipeline-aware ML Fairness: A Research Agenda for Developing Practical Guidelines and Tools
• Achieve Fairness without Demographics for Dermatological Disease Diagnosis
• Benign Shortcut for Debiasing: Fair Visual Recognition via Intervention with Shortcut Features
• Fairness without Demographics through Learning Graph of Gradients
• GroupMixNorm Layer for Learning Fair Models