Directed Structural Adaptation to Overcome Statistical Conflicts and Enable Continual Learning

TL;DR

This paper introduces DIRAD, a structural adaptation method for neural networks that dynamically complexifies to overcome statistical conflicts, and PREVAL, a framework for continual learning that prevents catastrophic forgetting without task labels.

Contribution

The paper presents DIRAD for directed, flexible network growth and PREVAL for task detection and continual learning without task labels, advancing adaptive neural network design.

Findings

DIRAD grows networks efficiently with high performance

PREVAL detects and distinguishes new and old tasks in continual learning

Networks using DIRAD and PREVAL outperform fixed topology models

Abstract

Adaptive networks today rely on overparameterized fixed topologies that cannot break through the statistical conflicts they encounter in the data they are exposed to, and are prone to "catastrophic forgetting" as the network attempts to reuse the existing structures to learn new task. We propose a structural adaptation method, DIRAD, that can complexify as needed and in a directed manner without being limited by statistical conflicts within a dataset. We then extend this method and present the PREVAL framework, designed to prevent "catastrophic forgetting" in continual learning by detection of new data and assigning encountered data to suitable models adapted to process them, without needing task labels anywhere in the workflow. We show the reliability of the DIRAD in growing a network with high performance and orders-of-magnitude simpler than fixed topology networks; and demonstrate the proof-of-concept operation of PREVAL, in which continual adaptation to new tasks is observed while being able to detect and discern previously-encountered tasks.