ARE NEURAL NETS MODULAR? INSPECTING FUNCTIONAL MODULARITY THROUGH DIFFERENTIABLE WEIGHT MASKS

Róbert Csordás, Sjoerd van Steenkiste, Jürgen Schmidhuber

Research output: Contribution to conferencePaperpeer-review

21 Scopus citations

Abstract

Neural networks (NNs) whose subnetworks implement reusable functions are expected to offer numerous advantages, including compositionality through efficient recombination of functional building blocks, interpretability, preventing catastrophic interference, etc. Understanding if and how NNs are modular could provide insights into how to improve them. Current inspection methods, however, fail to link modules to their functionality. In this paper, we present a novel method based on learning binary weight masks to identify individual weights and subnets responsible for specific functions. Using this powerful tool, we contribute an extensive study of emerging modularity in NNs that covers several standard architectures and datasets. We demonstrate how common NNs fail to reuse submodules and offer new insights into the related issue of systematic generalization on language tasks.

Original languageEnglish (US)
StatePublished - 2021
Event9th International Conference on Learning Representations, ICLR 2021 - Virtual, Online
Duration: May 3 2021May 7 2021

Conference

Conference9th International Conference on Learning Representations, ICLR 2021
CityVirtual, Online
Period05/3/2105/7/21

ASJC Scopus subject areas

  • Language and Linguistics
  • Computer Science Applications
  • Education
  • Linguistics and Language

Fingerprint

Dive into the research topics of 'ARE NEURAL NETS MODULAR? INSPECTING FUNCTIONAL MODULARITY THROUGH DIFFERENTIABLE WEIGHT MASKS'. Together they form a unique fingerprint.

Cite this