Distributed Learning with Adversarial Gradient Perturbations

Nawapon Sangsiri, Yufei Tao

Privacy concerns in distributed learning often lead clients to return intentionally altered gradient information. We consider the problem of learning convex and $L$-smooth function…