Sebra: Debiasing Through Self-Guided Bias Ranking

2025-01-30

AI Summary

This paper addresses training bias in unsupervised machine learning models caused by spurious correlations. To tackle this, we propose Sebra—a novel framework for end-to-end unsupervised debiasing. Its core contribution is the first identification of a local symmetry in Empirical Risk Minimization (ERM): ease of learning is negatively correlated with spuriousness. Leveraging this insight, Sebra introduces a self-guided sample ranking mechanism that automatically orders samples by intra-class fine-grained spuriousness—without requiring human annotations. Furthermore, it integrates dynamic bias correction with multi-source-aware contrastive learning to jointly suppress spurious features and enhance invariant representations. Extensive experiments demonstrate that Sebra significantly outperforms existing unsupervised debiasing methods on UrbanCars, BAR, CelebA, and ImageNet-1K. To foster reproducibility, we release our code, pre-trained models, and experimental logs.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Bias, Fairness & Privacy
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Ranking samples by fine-grained estimates of spuriosity (the degree to which spurious cues are present) has recently been shown to significantly benefit bias mitigation, over the traditional binary biased- extit{vs}-unbiased partitioning of train sets. However, this spuriosity ranking comes with the requirement of human supervision. In this paper, we propose a debiasing framework based on our novel ul{Se}lf-Guided ul{B}ias ul{Ra}nking (emph{Sebra}), that mitigates biases (spurious correlations) via an automatic ranking of data points by spuriosity within their respective classes. Sebra leverages a key local symmetry in Empirical Risk Minimization (ERM) training -- the ease of learning a sample via ERM inversely correlates with its spuriousity; the fewer spurious correlations a sample exhibits, the harder it is to learn, and vice versa. However, globally across iterations, ERM tends to deviate from this symmetry. Sebra dynamically steers ERM to correct this deviation, facilitating the sequential learning of attributes in increasing order of difficulty, ie, decreasing order of spuriosity. As a result, the sequence in which Sebra learns samples naturally provides spuriousity rankings. We use the resulting fine-grained bias characterization in a contrastive learning framework to mitigate biases from multiple sources. Extensive experiments show that Sebra consistently outperforms previous state-of-the-art unsupervised debiasing techniques across multiple standard benchmarks, including UrbanCars, BAR, CelebA, and ImageNet-1K. Code, pre-trained models, and training logs are available at https://kadarsh22.github.io/sebra_iclr25/.

Problem

• Automatic Debiasing
• Machine Learning Models
• Misleading Information

Innovation

• Sebra
• Bias Reduction
• Self-Guided Ranking