CPR: Classifier-Projection Regularization for Continual Learning

TL;DR

This paper introduces Classifier-Projection Regularization (CPR), a simple method that enhances continual learning by promoting wide local minima and reducing forgetting, applicable to existing regularization-based approaches.

Contribution

The paper proposes a novel regularization technique, CPR, that maximizes classifier output entropy to improve continual learning performance and mitigate catastrophic forgetting.

Findings

CPR improves accuracy on image recognition benchmarks.

CPR enhances plasticity and stability in continual learning.

CPR promotes wide local minima in neural networks.

Abstract

We propose a general, yet simple patch that can be applied to existing regularization-based continual learning methods called classifier-projection regularization (CPR). Inspired by both recent results on neural networks with wide local minima and information theory, CPR adds an additional regularization term that maximizes the entropy of a classifier's output probability. We demonstrate that this additional term can be interpreted as a projection of the conditional probability given by a classifier's output to the uniform distribution. By applying the Pythagorean theorem for KL divergence, we then prove that this projection may (in theory) improve the performance of continual learning methods. In our extensive experimental results, we apply CPR to several state-of-the-art regularization-based continual learning methods and benchmark performance on popular image recognition datasets. Our results demonstrate that CPR indeed promotes a wide local minima and significantly improves both accuracy and plasticity while simultaneously mitigating the catastrophic forgetting of baseline continual learning methods. The codes and scripts for this work are available at https://github.com/csm9493/CPR_CL.