Fetching the paper…

DRACO: Byzantine-resilient Distributed Training via Redundant Gradients · Around