no code implementations • 19 Feb 2019 • Chris Mesterharm, Rauf Izmailov, Scott Alexander, Simon Tsang
In this paper, we consider batch supervised learning where an adversary is allowed to corrupt instances with arbitrarily large noise.
