Speaker: Tianyi Chen

Time and Date: 9:30-10:30 am, November 15, 2018

Place: Room B415 of Computing Center Building, Handan Campus, Fudan University

Abstract:

Internet-of-Things (IoT) envisions an intelligent infrastructure of networked smart devices offering task-specific monitoring and control services. To meet the low-latency and privacy-preserving requirements of IoT applications, machine learning tasks need to be performed on IoT devices. However, the unique features of IoT such as heterogeneity and massive number of devices prevent successful implementation of existing centralized learning schemes in IoT settings. In this context, the nascent field of distributed machine learning or federated learning explores training statistical models over massive number of IoT devices. This talk will highlight the key challenges in distributed machine learning, including high communication overhead, stragglers, and fault tolerance. By marrying systems-level considerations and advanced optimization techniques, we will introduce novel distributed machine learning schemes with order-of-magnitude speedups for solving supervised and reinforcement learning tasks on IoT devices. We corroborate impressive empirical results with theoretical guarantees to give further insight of our algorithmic design.

Biography:

Tianyi Chen received the B. Eng. degree in Communication Science and Engineering from Fudan University, and the M.Sc. degree in Electrical and Computer Engineering (ECE) from the University of Minnesota (UMN), in 2014 and 2016, respectively. Since July 2016, he has been working toward his Ph.D. degree at UMN. His research interests lie in machine learning and large-scale optimization with applications to communication systems, energy systems, and Internet-of-Things. He was in the Best Student Paper Award finalist of the Asilomar Conference on Signals, Systems, and Computers. He received the National Scholarship from China in 2013, UMN ECE Department Fellowship in 2014, and the UMN Doctoral Dissertation Fellowship in 2017.