Incorporating Neuro-Inspired Adaptability for Continual Learning in Artificial Intelligence

TL;DR

This paper introduces a neuro-inspired continual learning approach that balances memory stability and plasticity by modeling biological adaptive mechanisms, improving AI's ability to learn incrementally without catastrophic forgetting.

Contribution

It proposes a novel multi-learner architecture inspired by Drosophila learning systems that attenuates old memories to enhance plasticity and solution compatibility.

Findings

Outperforms synaptic regularization methods in task-incremental learning.

Theoretical validation supports improved memory attenuation and adaptability.

Empirical results demonstrate enhanced continual learning performance.

Abstract

Continual learning aims to empower artificial intelligence (AI) with strong adaptability to the real world. For this purpose, a desirable solution should properly balance memory stability with learning plasticity, and acquire sufficient compatibility to capture the observed distributions. Existing advances mainly focus on preserving memory stability to overcome catastrophic forgetting, but remain difficult to flexibly accommodate incremental changes as biological intelligence (BI) does. By modeling a robust Drosophila learning system that actively regulates forgetting with multiple learning modules, here we propose a generic approach that appropriately attenuates old memories in parameter distributions to improve learning plasticity, and accordingly coordinates a multi-learner architecture to ensure solution compatibility. Through extensive theoretical and empirical validation, our approach not only clearly enhances the performance of continual learning, especially over synaptic regularization methods in task-incremental settings, but also potentially advances the understanding of neurological adaptive mechanisms, serving as a novel paradigm to progress AI and BI together.