Keynotes at MLSys’23 Workshops!

I will be giving two keynote talks next week at MLSys workshops on "Federated Learning Systems" and "Resource-Constrained Learning in Wireless Networks” to discuss federated training of large models over resource-constrained edge nodes. 

There are three promising approaches to this problem: “low-rank sub-model training”, “sparse training”, and “knowledge transfer from generative pre-trained models”. I will discuss the pros and cons of each approach and our recent results that demonstrate their promise.