Sign in

Attacking Byzantine Robust Aggregation in High Dimensions

By Sarthak Choudhary and others at
LogoNational University of Singapore
Training modern neural networks or models typically requires averaging over a sample of high-dimensional vectors. Poisoning attacks can skew or bias the average vectors used to train the model, forcing the model to learn specific patterns or avoid learning anything useful. Byzantine robust aggregation is a principled algorithmic defense against... Show more
April 19, 2024
=
0
Loading PDF…
Loading full text...
Similar articles
Loading recommendations...
=
0
x1
Attacking Byzantine Robust Aggregation in High Dimensions
Click on play to start listening