Yi-Rui Yang

On the tension between Byzantine robustness and no-attack accuracy in distributed learning