<4rd Digital Universe Colloquium>

-Title: Recent Trends of Federated Learning 

- Speker: Yoon, Sung Whan (Host: Kim, Ue Hwan)

               (Assistant Professor of the UNIST)

- Date/Time: October 10, 2024 (Thursday) 16:00–17:00

- Location: Building A(N4), Room 227

Below is an abstract of the talk:

In this talk, we study the recent trends of federated learning (FL), which draws great attention as a key enabler for a data-secured distributed learning framework. From the pioneering work of the baseline FL called FedAvg, many researchers have dedicated themselves to pushing it forward to be a high-performing, efficient, and reliable distributed learning framework by overcoming various challenges, including fundamental theories, data heterogeneity, data secrecy, etc. We briefly address the current interests of FL in research society and introduce the recent works done by our research group. Specifically, our group focuses on handling data heterogeneity issues and the theoretical foundations of FL. Finally, the future direction for resolving the remaining challenges of FL will be discussed.

**Students interested in the lecture can attend even if they haven't registered, provided there are available seats in the classroom.