"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis cover image

Dodging Latent Space Detectors: Obfuscated Activation Attacks with Luke, Erik, and Scott.

"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis

00:00

Mastering Obfuscation in Machine Learning

This chapter explores the dynamics of obfuscation attacks in machine learning, detailing how attackers optimize loss functions to make models produce harmful outputs while evading detection. It discusses the complexities involved in training both attack models and monitoring systems simultaneously, as well as the manipulation of training datasets to enhance attack effectiveness. The chapter also examines various attack strategies, including data poisoning and the implications of latent space manipulation, highlighting the evolving cat-and-mouse game between attackers and defenders.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app