Look-Ahead Selective Plasticity for Continual Learning of Visual Tasks

TL;DR

This paper introduces a novel continual learning method inspired by brain event models, using contrastive loss to identify and retain important network parameters during task transitions, leading to improved performance on vision benchmarks.

Contribution

It proposes a new mechanism that leverages contrastive loss at task boundaries to selectively preserve parameters, enhancing continual learning without relying on previous task data.

Findings

Achieves state-of-the-art results on CIFAR10 and TinyImagenet

Effective in task-incremental, class-incremental, and domain-incremental scenarios

Reduces catastrophic forgetting by selective parameter retention

Abstract

Contrastive representation learning has emerged as a promising technique for continual learning as it can learn representations that are robust to catastrophic forgetting and generalize well to unseen future tasks. Previous work in continual learning has addressed forgetting by using previous task data and trained models. Inspired by event models created and updated in the brain, we propose a new mechanism that takes place during task boundaries, i.e., when one task finishes and another starts. By observing the redundancy-inducing ability of contrastive loss on the output of a neural network, our method leverages the first few samples of the new task to identify and retain parameters contributing most to the transfer ability of the neural network, freeing up the remaining parts of the network to learn new features. We evaluate the proposed methods on benchmark computer vision datasets including CIFAR10 and TinyImagenet and demonstrate state-of-the-art performance in the task-incremental, class-incremental, and domain-incremental continual learning scenarios.