Balanced Gradient Sample Retrieval for Enhanced Knowledge Retention in Proxy-based Continual Learning

TL;DR

This paper introduces a novel sample retrieval method for continual learning that balances gradient-conflicting and gradient-aligned samples to better retain knowledge and reduce forgetting in neural networks.

Contribution

It proposes a new retrieval strategy leveraging both conflicting and aligned samples within a supervised contrastive learning framework for improved continual learning.

Findings

Outperforms existing methods in mitigating forgetting.

Achieves state-of-the-art results on popular benchmarks.

Enhances knowledge retention by balancing sample types.

Abstract

Continual learning in deep neural networks often suffers from catastrophic forgetting, where representations for previous tasks are overwritten during subsequent training. We propose a novel sample retrieval strategy from the memory buffer that leverages both gradient-conflicting and gradient-aligned samples to effectively retain knowledge about past tasks within a supervised contrastive learning framework. Gradient-conflicting samples are selected for their potential to reduce interference by re-aligning gradients, thereby preserving past task knowledge. Meanwhile, gradient-aligned samples are incorporated to reinforce stable, shared representations across tasks. By balancing gradient correction from conflicting samples with alignment reinforcement from aligned ones, our approach increases the diversity among retrieved instances and achieves superior alignment in parameter space, significantly enhancing knowledge retention and mitigating proxy drift. Empirical results demonstrate that using both sample types outperforms methods relying solely on one sample type or random retrieval. Experiments on popular continual learning benchmarks in computer vision validate our method's state-of-the-art performance in mitigating forgetting while maintaining competitive accuracy on new tasks.