KEY NOTES

Dietterich

Thomas Dietterich

Oregon State University

Talk title: Anomaly Detection for Out-of-Distribution and Open Set Detection

Abstract: Standard supervised classification assumes a closed iid world: Test data come from the same distribution as the training data, and they contain only the same classes as the training data. How can we detect when these assumptions are violated? One approach is to train an anomaly detector that checks each test query to see if it falls within the region of competence of the classifier. If the query is anomalous because of a change of distribution and particularly if the change introduces a novel class, the anomaly detector should reject the query. This talk will review the current state of the art in this area with a focus on deep anomaly detection. A key challenge is that whereas in standard anomaly detection the engineered feature space can provide a useful distance measure, in the representations learned by deep learning, distances are much less reliable.

Short bio: Dr. Dietterich (AB Oberlin College 1977; MS University of Illinois 1979; PhD Stanford University 1984) is Distinguished Professor Emeritus in the School of Electrical Engineering and Computer Science at Oregon State University. Dietterich is one of the pioneers of the field of Machine Learning and has authored more than 220 refereed publications and two books. His current research topics include robust artificial intelligence, robust human-AI systems, and applications in sustainability. Dietterich has devoted many years of service to the research community. He is a former President of the Association for the Advancement of Artificial Intelligence, and the founding president of the International Machine Learning Society. Other major roles include Executive Editor of the journal Machine Learning, co-founder of the Journal for Machine Learning Research, and program chair of AAAI 1990 and NIPS 2000. He currently serves as one of the moderators for the cs.LG category on arXiv.

Yiran

Yiran Chen

Duke University

Talk title: Efficient and Reliable Deep Learning at Scale

Abstract: The rapid growth of modern neural network (NN) models’ scale generates ever-increasing demands for high computing power of artificial intelligence (AI) systems. Many specialized computing devices have been also deployed in the AI systems, forming a truly application-driven heterogeneous computing platform. This talk discusses the importance of computing system and paradigm designs in AI applications. We first discuss the design philosophy of heterogeneous AI computing systems, and then present several hardware friendly efficient NN model design techniques. We also extend our discussions to distributed systems and federated learning, and briefly introduce the automation of the co-design flow, e.g., neural architecture search. A research roadmap of our group in the relevant topics is given at the end of the talk.

Short Bio: Yiran Chen received B.S (1998) and M.S. (2001) from Tsinghua University and Ph.D. (2005) from Purdue University. After five years in industry, he joined University of Pittsburgh in 2010 as Assistant Professor and then was promoted to Associate Professor with tenure in 2014, holding Bicentennial Alumni Faculty Fellow. He is now the Professor of the Department of Electrical and Computer Engineering at Duke University and serving as the director of the NSF AI Institute for Edge Computing Leveraging the Next-generation Networks (Athena) and the NSF Industry–University Cooperative Research Center (IUCRC) for Alternative Sustainable and Intelligent Computing (ASIC), and the co-director of Duke Center for Computational Evolutionary Intelligence (CEI). His group focuses on the research of new memory and storage systems, machine learning and neuromorphic computing, and mobile computing systems. Dr. Chen has published 1 book and about 500 technical publications and has been granted 96 US patents. He has served as the associate editor of a dozen international academic transactions/journals and served on the technical and organization committees of more than 60 international conferences. He is now serving as the Editor-in-Chief of the IEEE Circuits and Systems Magazine. He received seven best paper awards, one best poster award, and fifteen best paper nominations from international conferences and workshops. He received many professional awards and is the distinguished lecturer of IEEE CEDA (2018-2021). He is a Fellow of the ACM and IEEE and now serves as the chair of ACM SIGDA.

Sorin

Sorin Draghici

Wayne State University

Talk title: Using graph methods to understand diseases and repurpose drugs

Abstract: This talk will discuss some advanced bioinformatics methods and software that can be used to greatly facility the analysis and understanding of high-throughput omics data. Issues will include drawbacks of commonly used approaches for pathway analysis, a method comparison and benchmarking results, and other issues related to data analysis. The seminar will include live demonstrations on several data sets aiming at understanding the disease and identifying drugs suitable for repurposing in COVID-19.

Short Bio: Dr. Draghici is a Professor in the Department of Computer Science, and the head of the Intelligent Systems and Bioinformatics Laboratory at Wayne State University. He also holds a joint appointment in the Department Obstetrics and Gynecology and is an Associate Dean in Wayne State University's College of Engineering. Dr. Draghici is a senior member of IEEE, and an editor of IEEE/ACM Transactions on Computational Biology and Bioinformatics, Protocols in Bioinformatics, Discoveries Journals, Journal of Biomedicine and Biotechnology, and International Journal of Functional Informatics and Personalized Medicine. His publications include two books (”Data Analysis Tools for DNA Microarrays and Statistics” and ”Data Analysis for Microarrays using R”), 8 book chapters, and over 150 peer-reviewed journal and conference publications which gathered over 12,000 citations to date.