Two exciting speakers will be delivering the A2IR2 webinars for August and September. The August seminar will be presented by Prof. Sanmi Koyejo. Sanmi Koyejo is an Assistant Professor in the Department of Computer Science at the University of Illinois at Urbana-Champaign. Koyejo’s research interests are in developing the principles and practice of trustworthy machine learning. Additionally, Koyejo focuses on applications to neuroscience and healthcare. Koyejo completed his Ph.D. in Electrical Engineering at the University of Texas at Austin, advised by Joydeep Ghosh, and completed postdoctoral research at Stanford University. His postdoctoral research was primarily with Russell A. Poldrack and Pradeep Ravikumar. Koyejo has been the recipient of several awards, including a best paper award from the conference on uncertainty in artificial intelligence (UAI), a Skip Ellis Early Career Award, a Sloan Fellowship, a Kavli Fellowship, an IJCAI early career spotlight, and a trainee award from the Organization for Human Brain Mapping (OHBM). Koyejo serves as the president of the Black in AI organization.
He will be delivering a presentation on Fault-tolerant federated and distributed machine learning. Machine learning (ML) models are routinely trained and deployed among devices that are susceptible to hardware/software/communication errors and/or security concerns. Examples include geo-distributed datacenters with non-negligible communication latency, groups of mobile devices or Internet of Things (IoT), and volunteer ML computing. For such settings, distributed training typically consists of separate updates interleaved with aggregation. To this end, I will outline novel aggregation schemes for fault-tolerant federated learning and distributed training via stochastic gradient descent. The proposed aggregation schemes are shown to be provably robust to worst-case errors from a large fraction of arbitrarily malicious workers (aka Byzantine errors), with minimal effect on convergence rates. Empirical evaluation in a variety of real-world setting further highlights the performance of the proposed aggregation strategies.
We will follow this up with a presentation in September by Prof. Igor Mezic of the Departments of Mechanical Engineering and Mathematics, the University of California, Santa Barbara. Prof. Mezic is best known for his work on operator theoretic methods, and he is largely credited for the re-emergence of the Koopman operator methods for nonlinear analysis.
Prof. Mezic holds a Ph.D. from the California Institute of Technology, an was a postdoctoral fellow at the Mathematics Institute of the University of Warwick, in the UK. He joined the faculty of UCSB in 1995, where he became a full professor in 2003. He has received numerous awards and honours for his research spanning physics, mathematics and engineering. He is a Fellow of the American Physical Society and the Society for Industrial and Applied Mathematics. He is also the first ever Honoris Professor of the University of Rijeka, Croatia. He holds numerous US patents on which at least three technology startups have been founded.
Prof. Mezic’s seminar will be on the subject of Koopman theoretic methods and will be delivered with the express goal of stimulating the creation of a large, active community of scholars in Africa developing and using Koopman theoretic methods. The title and abstract for Prof. Mezic’s presentation will be circulated within the next few weeks.
