Machine learning has an alarming threat: undetectable backdoors

-

This article is part of our coverage of the latest in AI research. If an adversary gives you a machine learning model and secretly plants a malicious backdoor in it, what are the chances that you can discover it? Very little, according to a new paper by researchers at UC Berkeley, MIT, and the Institute of Advanced Study. The security of machine learning is becoming increasingly critical as ML models find their way into a growing number of applications. The new study focuses on the security threats of delegating the training and development of machine learning models to third parties and service providers. With…

This story continues at The Next Web

from The Next Web https://ift.tt/3QWXt7p

Comments

Post a Comment

Popular posts from this blog

Quantum physicists say time travelers don’t have to worry about the butterfly effect

-
Image
What if I told you all your favorite time-travel films and books were actually created by big tech in order to wrest control of the time-travel industry from the proletariat? Think about it. Back to the Future, The Terminator, The Time Machine, all of these stories share a central theme where traveling through time is a dangerous proposition that could destroy the very fabric of our reality. It’s called the butterfly effect. The big idea is that you’d step out of your time travel machine and accidentally step on a bug. Because this bug doesn’t exist… maybe a frog goes… This story continues at The Next Web from The Next Web https://ift.tt/2XfjPI6
Read more