Bayesian Gated Non-Negative Contrastive Learning

2026-05-27

AI Summary

This work addresses the issue of representation entanglement in contrastive learning, which arises from deterministic similarity metrics and is exacerbated in compositional scenarios with co-occurring background features, leading to optimization conflicts. To mitigate this, the paper proposes BayesNCL, the first approach integrating Bayesian variational inference with a probabilistic gating mechanism. By employing a sparse Bernoulli prior, BayesNCL dynamically filters out task-irrelevant yet frequently co-occurring features, framing feature selection as a variational inference problem to achieve semantic disentanglement. The method effectively alleviates gradient oscillations caused by shared features in positive and negative sample pairs. Experiments demonstrate a 142.1% improvement in semantic consistency over the strongest baseline on ImageNet-100, while maintaining competitive downstream task performance.

Abstract

While Contrastive Learning (CL) has revolutionized self-supervised representation learning, its latent representations remain highly entangled and opaque, limiting their interpretability in safety-critical applications. We identify that a fundamental cause of this entanglement is the reliance on deterministic similarity measures, which treat all feature dimensions equally. In compositional scenes, this creates an Optimization Conflict: common background features, such as, "blue sky", are encouraged to align in positive pairs but simultaneously repelled in negative pairs, causing gradient oscillations that hinder precise semantic disentanglement. To address this, we propose BayesNCL (Bayesian Gated Non-Negative Contrastive Learning). Unlike standard approaches, BayesNCL introduces a probabilistic gating mechanism that dynamically filters out task-irrelevant, high-frequency common features while selectively retaining discriminative semantics. By formalizing feature selection as a variational inference problem with a sparse Bernoulli prior, our method effectively resolves the optimization conflict. Empirical experimental results on Imagenet-100 demonstrate that BayesNCL achieves a remarkable 142.1% improvement in semantic consistency compared to state-of-the-art baselines, yielding highly interpretable representations without compromising downstream task performance. Code is available at https://github.com/Cui-Peng-624/BayesNCL.

Problem

• Contrastive Learning
• Representation Disentanglement
• Optimization Conflict
• Interpretability
• Self-supervised Learning

Innovation

• Bayesian Gating
• Contrastive Learning
• Semantic Disentanglement
• Variational Inference
• Non-Negative Learning