Can invisible physical events easily fool spiking neural networks?

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

Abstract

Event-based dynamic vision sensors, which generate sparse spike-based outputs, are ideal for low-power applications. Spiking Neural Networks are designed to process this data efficiently on asynchronous neuromorphic hardware. As event-based vision advances, understanding the vulnerability of Spiking Neural Networks to physical adversarial attacks becomes crucial. This work introduces a novel light-based adversarial attack on neuromorphic vision. We exploit undetectable optical events, specifically designed light pulses, to disrupt the temporal dynamics of event-based sensors. Our method demonstrates how these physical attacks can be tailored to the event-based data's discrete and sparse nature while achieving high success rates.

Original languageEnglish
Title of host publicationArtificial Intelligence for Security and Defence Applications III
EditorsHugo J. Kuijf, Radhakrishna Prabhu, Yitzhak Yitzhaky
PublisherSPIE
ISBN (Electronic)9781510692978
DOIs
StatePublished - 28 Oct 2025
Event3rd Artificial Intelligence for Security and Defence Applications - Madrid, Spain
Duration: 16 Sep 202518 Sep 2025

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume13679
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Conference

Conference3rd Artificial Intelligence for Security and Defence Applications
Country/TerritorySpain
CityMadrid
Period16/09/2518/09/25

Keywords

  • Dynamicvision sensors
  • Neuromorphic vision
  • Optical perturbations
  • Physical adversarial events
  • Robust AI
  • Spiking neural networks

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Instrumentation
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Can invisible physical events easily fool spiking neural networks?'. Together they form a unique fingerprint.

Cite this