‘Triggerless’ backdoors can infect machine learning models without leaving a trace – research
![‘Triggerless’ backdoors can infect machine learning models without leaving a trace – research](https://i0.wp.com/www.haktechs.com/wp-content/uploads/2020/11/‘Triggerless-backdoors-can-infect-machine-learning-models-without-leaving-a.jpg?fit=506%2C254&ssl=1)
Malicious actors can induce AI devices to act erratically with out utilizing visible ‘triggers’, scientists at the Germany-primarily based CISPA Helmholtz Centre for Information Protection have identified.