Skip to main navigation Skip to search Skip to main content

Optimal Activation Functions via Causal Wyner-Ziv Coding

  • University of Bern

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We propose a new method for deriving optimal activation functions that relate the strength of a neuron’s stimulus to its firing intensity. Our method is based on an information-theoretic model of neuronal communication and yields activation functions that provably maximize the communication efficiency at a prespecified fidelity. Crucially, we show that such activation functions exist (under a mild assumption on the required fidelity) and provide an example thereof.

Original languageEnglish
Title of host publication2024 IEEE International Symposium on Information Theory Workshops, ISIT-W 2024
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350348446
DOIs
Publication statusPublished - 1 Jan 2024
Event2024 IEEE International Symposium on Information Theory Workshops, ISIT-W 2024 - Athens, Greece
Duration: 7 Jul 2024 → …

Publication series

Name2024 IEEE International Symposium on Information Theory Workshops, ISIT-W 2024

Conference

Conference2024 IEEE International Symposium on Information Theory Workshops, ISIT-W 2024
Country/TerritoryGreece
CityAthens
Period7/07/24 → …

Fingerprint

Dive into the research topics of 'Optimal Activation Functions via Causal Wyner-Ziv Coding'. Together they form a unique fingerprint.

Cite this