Machine Learning

We designed a decentralized multi-task learning framework and a novel optimization algorithm to collaboratively train models over distributed devices by only sharing their gradient periodically