A Neural Network Model of Continual Learning with Cognitive Control

TL;DR

This paper demonstrates that neural networks with a cognitive control mechanism can effectively prevent catastrophic forgetting in continual learning, mirroring human advantages of blocking and offering insights into neural representations.

Contribution

It introduces a cognitive control mechanism into neural networks that mitigates catastrophic forgetting and explains the human-like benefit of blocking in learning.

Findings

Networks with cognitive control avoid catastrophic forgetting.

Blocking can outperform interleaving with active maintenance.

Analysis reveals map-like representations related to control mechanisms.

Abstract

Neural networks struggle in continual learning settings from catastrophic forgetting: when trials are blocked, new learning can overwrite the learning from previous blocks. Humans learn effectively in these settings, in some cases even showing an advantage of blocking, suggesting the brain contains mechanisms to overcome this problem. Here, we build on previous work and show that neural networks equipped with a mechanism for cognitive control do not exhibit catastrophic forgetting when trials are blocked. We further show an advantage of blocking over interleaving when there is a bias for active maintenance in the control signal, implying a tradeoff between maintenance and the strength of control. Analyses of map-like representations learned by the networks provided additional insights into these mechanisms. Our work highlights the potential of cognitive control to aid continual learning in neural networks, and offers an explanation for the advantage of blocking that has been observed in humans.