Neural Network Plasticity and Loss Sharpness

TL;DR

This paper investigates the relationship between neural network plasticity and loss landscape sharpness in continual learning, finding that sharpness regularization techniques do not effectively reduce plasticity loss in non-stationary environments.

Contribution

The study examines the effectiveness of sharpness regularization methods in mitigating plasticity loss during continual learning, revealing their limited impact.

Findings

Sharpness regularization techniques do not significantly reduce plasticity loss.

Plasticity loss is highly related to loss landscape sharpness in non-stationary RL.

Regularization methods aimed at smooth minima may not improve adaptation in continual learning.

Abstract

In recent years, continual learning, a prediction setting in which the problem environment may evolve over time, has become an increasingly popular research field due to the framework's gearing towards complex, non-stationary objectives. Learning such objectives requires plasticity, or the ability of a neural network to adapt its predictions to a different task. Recent findings indicate that plasticity loss on new tasks is highly related to loss landscape sharpness in non-stationary RL frameworks. We explore the usage of sharpness regularization techniques, which seek out smooth minima and have been touted for their generalization capabilities in vanilla prediction settings, in efforts to combat plasticity loss. Our findings indicate that such techniques have no significant effect on reducing plasticity loss.