M365 Research

Pioneering advancements in AI and systems

新闻与深度文章

This diagram shows a payload exchange between a server, inside Worker 0, and clients that live inside Workers 2 and 3. First, the server pushes the central ML model plus the clients’ data to Workers 2 and 3. Then, each client trains the model with their local data. Finally, the clients send the pseudo-gradients of this new model back to the server for aggregation and the creation of a new global model.
微软研究院博客

| Dimitrios Dimitriadis, Mirian Hipolito Garcia, Daniel Eduardo Madrigal Diaz, Andre Manoel, 和 Robert Sim

Federated learning has become a major area of machine learning (ML) research in recent years due to its versatility in training complex models over massive amounts of data without the need to share that data with a centralized entity. However,…

Graphic shows framework of Privacy Preserving Machine Learning
微软研究院博客

| Victor Ruehle, Robert Sim, Sergey Yekhanin, Nishanth Chandran, Melissa Chase, Daniel Jones, Kim Laine, Boris Köpf, Jaime Teevan, Jim Kleewein, 和 Saravan Rajmohan

Machine learning (ML) offers tremendous opportunities to increase productivity. However, ML systems are only as good as the quality of the data that informs the training of ML models. And training ML models requires a significant amount of data, more…