ISQED'19 Keynotes
- Check out selected keynotes from previous years.
Wednesday & Thursday March 6-7
Adversarial attacks on Security and Privacy of Machine Learning Systems
Sandip KunduSandip Kundu - Program Director, National Science Foundation (NSF)
As applications of machine learning (ML) have become commonplace in healthcare, security, finance and many other mission critical systems, the security risk of machine-learning systems is emerging as a concern. Machine learning applications evolve through multiple stages including data collection, preparation, labeling, model training, testing and deployment. Malicious actors can impact the reliability and dependability of machine learning system by exploiting vulnerability at any of these stages. In this talk, we present a taxonomy for attack that categorizes an attack based on three fundamental pillars of information security, namely, confidentiality, integrity, and availability. In confidentiality violation, an adversary uses the ML responses to infer the model parameters, or the secret information used in learning process. In an integrity violation, the attacker causes to allow harmful instances to slip through the ML model as false negatives. In an availability violation, the attacker creates a denial of service event. We conclude the talk with various approaches for defending against adversarial attack on ML system.
About Sandip Kundu
Sandip Kundu is a Program Director at the National Science Foundation in the CNS division within the CISE directorate. He is serving in this position on leave from the University of Massachusetts at Amherst, where he is a professor in Electrical and Computer Engineering Department. Kundu began his career at IBM Research as a Research Staff Member; then worked at Intel Corporation as a Principal Engineer before joining UMass Amherst as a professor in 2005. He has published over 250 research papers in VLSI design and test, holds several key patents including ultra-drowsy sleep mode in processors, and has given more than a dozen tutorials at various conferences. He is a Fellow of the IEEE, Fellow of the Japan Society for Promotion of Science (JSPS), Senior International Scientist of the Chinese Academy of Sciences and was a Distinguished Visitor of the IEEE Computer Society. He is currently an Associate Editor of the IEEE Transactions on Dependable and Secure Computing. Previously, he has served as an Associate Editor of the IEEE Transactions on Computers, IEEE Transactions on VLSI Systems and ACM Transactions on Design Automation of Electronic Systems. He has been Technical Program Chair/General Chair of multiple conferences including ICCD, ATS, ISVLSI, DFTS and VLSI Design Conference.
Machine Learning is Changing the Game for Variability and Characterization and will soon help Analog and Digital Verification
Amit GuptaAmit Gupta - General Manager of the IC Verification Solutions Solido division of Mentor, a Siemens Business
The Golden Age of machine learning is upon EDA. Over the past four years, we have seen large EDA suppliers and customers grow their internal ML teams and strategies, and ML research projects are emerging in all areas of EDA. But, we have not yet seen much of this investment convert into real production flows and work. This talk reviews a set of challenges that make it difficult to bring ML solutions to production for semiconductor design, and discusses approaches for solving them. We will discuss how these approaches are already benefiting variation-aware design and characterization flows, and the broader applicability to analog and digital verification.
About Amit Gupta
Amit Gupta is General Manager of the IC Verification Solutions Solido division of Mentor, a Siemens Business. Previously, he founded Solido Design Automation Inc. in 2005 and served as its President and CEO until its acquisition by Mentor in 2017. Solido is a leader in machine learning variation-aware design and characterization software. In 1999, he founded Analog Design Automation Inc. (ADA), and served as its President, CEO and VP of Business Development until it was acquired by Synopsys in 2004. ADA was a leader in analog optimization software. He has previously served as a Director of the Electronic Design Automation Consortium. Amit holds degrees in both Electrical Engineering and Computer Science with great Distinction from the University of Saskatchewan, and was awarded the 2005 outstanding alumni award for significant accomplishments since graduation.