Unbiased and Efficient Self-Supervised Incremental Contrastive Learning

TL;DR

This paper introduces a novel self-supervised incremental contrastive learning framework that reduces bias and improves efficiency by estimating noise distribution changes and adaptively learning the learning rate, enabling faster convergence without retraining.

Contribution

It proposes a new Incremental InfoNCE loss and a meta-optimization mechanism with deep reinforced learning rate learning for unbiased, efficient incremental contrastive learning.

Findings

Achieves up to 16.7x training speedup

Attains 16.8x faster convergence

Maintains competitive performance

Abstract

Contrastive Learning (CL) has been proved to be a powerful self-supervised approach for a wide range of domains, including computer vision and graph representation learning. However, the incremental learning issue of CL has rarely been studied, which brings the limitation in applying it to real-world applications. Contrastive learning identifies the samples with the negative ones from the noise distribution that changes in the incremental scenarios. Therefore, only fitting the change of data without noise distribution causes bias, and directly retraining results in low efficiency. To bridge this research gap, we propose a self-supervised Incremental Contrastive Learning (ICL) framework consisting of (i) a novel Incremental InfoNCE (NCE-II) loss function by estimating the change of noise distribution for old data to guarantee no bias with respect to the retraining, (ii) a meta-optimization with deep reinforced Learning Rate Learning (LRL) mechanism which can adaptively learn the learning rate according to the status of the training processes and achieve fast convergence which is critical for incremental learning. Theoretically, the proposed ICL is equivalent to retraining, which is based on solid mathematical derivation. In practice, extensive experiments in different domains demonstrate that, without retraining a new model, ICL achieves up to 16.7x training speedup and 16.8x faster convergence with competitive results.