Hypnopaedia-Aware Machine Unlearning via Psychometrics of Artificial Mental Imagery

Hypnopaedia-Aware Machine Unlearning via Psychometrics of Artificial Mental Imagery

Neural backdoors represent insidious cybersecurity loopholes that render learning machinery vulnerable to unauthorised manipulations, potentially enabling the weaponisation of artificial intelligence with catastrophic consequences. A backdoor attack involves the clandestine infiltration of a trigger...

Full description

Saved in:
Bibliographic Details
Main Authors: Ching-Chun Chang, Kai Gao, Shuying Xu, Anastasia Kordoni, Christopher Leckie, Isao Echizen
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Cybersecurity
machine unlearning
neural backdoors
psychometrics
reverse engineering
Online Access:https://ieeexplore.ieee.org/document/11025476/
Tags: Add Tag
No Tags, Be the first to tag this record!

Internet

https://ieeexplore.ieee.org/document/11025476/

Similar Items