all InfoSec news
CorruptEncoder: Data Poisoning based Backdoor Attacks to Contrastive Learning. (arXiv:2211.08229v1 [cs.CR])
Nov. 16, 2022, 2:20 a.m. | Jinghuai Zhang, Hongbin Liu, Jinyuan Jia, Neil Zhenqiang Gong
cs.CR updates on arXiv.org arxiv.org
Contrastive learning (CL) pre-trains general-purpose encoders using an
unlabeled pre-training dataset, which consists of images (called single-modal
CL) or image-text pairs (called multi-modal CL). CL is vulnerable to data
poisoning based backdoor attacks (DPBAs), in which an attacker injects poisoned
inputs into the pre-training dataset so the encoder is backdoored. However,
existing DPBAs achieve limited effectiveness. In this work, we propose new
DPBAs called CorruptEncoder to CL. Our experiments show that CorruptEncoder
substantially outperforms existing DPBAs for both single-modal and …
More from arxiv.org / cs.CR updates on arXiv.org
Jobs in InfoSec / Cybersecurity
SOC 2 Manager, Audit and Certification
@ Deloitte | US and CA Multiple Locations
IT Security Manager
@ Teltonika | Vilnius/Kaunas, VL, LT
Security Officer - Part Time - Harrah's Gulf Coast
@ Caesars Entertainment | Biloxi, MS, United States
DevSecOps Full-stack Developer
@ Peraton | Fort Gordon, GA, United States
Cybersecurity Cooperation Lead
@ Peraton | Stuttgart, AE, United States
Cybersecurity Engineer - Malware & Forensics
@ ManTech | 201DU - Customer Site,Herndon, VA