Combating catastrophic forgetting with developmental compression

TL;DR

This paper introduces developmental compression, a scalable method to mitigate catastrophic forgetting in neural networks by gradually compressing specialized networks into a generalized model, validated on robot control tasks.

Contribution

It presents a novel developmental compression technique that reduces catastrophic forgetting without requiring domain knowledge, improving scalability over existing methods.

Findings

Reduces catastrophic forgetting in neural networks.

Produces generalized models without overt specialization.

Validated on robot control problem.

Abstract

Generally intelligent agents exhibit successful behavior across problems in several settings. Endemic in approaches to realize such intelligence in machines is catastrophic forgetting: sequential learning corrupts knowledge obtained earlier in the sequence, or tasks antagonistically compete for system resources. Methods for obviating catastrophic forgetting have sought to identify and preserve features of the system necessary to solve one problem when learning to solve another, or to enforce modularity such that minimally overlapping sub-functions contain task specific knowledge. While successful, both approaches scale poorly because they require larger architectures as the number of training instances grows, causing different parts of the system to specialize for separate subsets of the data. Here we present a method for addressing catastrophic forgetting called developmental compression. It exploits the mild impacts of developmental mutations to lessen adverse changes to previously-evolved capabilities and `compresses' specialized neural networks into a generalized one. In the absence of domain knowledge, developmental compression produces systems that avoid overt specialization, alleviating the need to engineer a bespoke system for every task permutation and suggesting better scalability than existing approaches. We validate this method on a robot control problem and hope to extend this approach to other machine learning domains in the future.