all InfoSec news
An Adversarial Robustness Perspective on the Topology of Neural Networks. (arXiv:2211.02675v1 [cs.LG])
Nov. 8, 2022, 2:20 a.m. | Morgane Goibert, Thomas Ricatte, Elvis Dohmatob
cs.CR updates on arXiv.org arxiv.org
In this paper, we investigate the impact of neural networks (NNs) topology on
adversarial robustness. Specifically, we study the graph produced when an input
traverses all the layers of a NN, and show that such graphs are different for
clean and adversarial inputs. We find that graphs from clean inputs are more
centralized around highway edges, whereas those from adversaries are more
diffuse, leveraging under-optimized edges. Through experiments on a variety of
datasets and architectures, we show that these under-optimized …
More from arxiv.org / cs.CR updates on arXiv.org
IDEA: Invariant Defense for Graph Adversarial Robustness
1 day, 1 hour ago |
arxiv.org
FairCMS: Cloud Media Sharing with Fair Copyright Protection
1 day, 1 hour ago |
arxiv.org
Jobs in InfoSec / Cybersecurity
SOC 2 Manager, Audit and Certification
@ Deloitte | US and CA Multiple Locations
Senior InfoSec Manager - Risk and Compliance
@ Federal Reserve System | Remote - Virginia
Security Analyst
@ Fortra | Mexico
Incident Responder
@ Babcock | Chester, GB, CH1 6ER
Vulnerability, Access & Inclusion Lead
@ Monzo | Cardiff, London or Remote (UK)
Information Security Analyst
@ Unissant | MD, USA