Year: 2021
Type: preprint
Abstract: In this paper we explore whether the fundamental tool of experimental psychology, the behavioral experiment, has the power to generate insight not only into humans and animals, but artificial systems too. We apply the techniques of experimental psychology to investigating catastrophic forgetting in neural networks. We present a series of controlled experiments with two-layer ReLU networks, and exploratory results revealing a new understanding of the behavior of catastrophic forgetting. Alongside our empirical findings, we demonstrate an alternative, behavior-first approach to investigating neural network phenomena.
Source: arXiv (Cornell University)
Authors: Bell, Samuel J., Lawrence, Neil D.
Institution: University of Cambridge
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3206011216