I have just delivered a keynote talk at ACL workshop on Federated Learning for Natural Language Processing (FL4NLP)

Secure, Scalable, and Efficient Federated Learning

Federated learning (FL) has emerged as a promising approach to enable decentralized machine learning directly at the edge, in order to enhance users' privacy, comply with regulations, and reduce development costs. In this talk, I will provide an overview of FL and highlight several key research directions in this area. In particular, I discuss four important research directions: (1) privacy and security guarantees of FL; (2) FL over resource-constrained edge nodes; (3) label scarcity and self-supervised FL; and (4) scalable system design for FL. In the second part of the talk, I will provide an overview of FedML, which is a machine learning platform that enables zero-code, lightweight, cross-platform, and provably secure federated learning and analytics. In particular, I highlight four key components of FedML platform: (1) an open-source community of more than 1k users; (2) a lightweight and cross-platform Edge AI SDK for deployment over GPUs, smartphones, and IoTs; (3) a user-friendly MLOps platform to simplify collaboration and real-world deployment; and (4) its diverse applications ecosystem (computer vision, natural language processing, data mining, and time-series forecasting).

The video recording will also be available soon: https://fl4nlp.github.io