Discrete Optimization in Machine Learning
With the premise that obtaining significantly better performance from pruning depends on accounting for the combined effect of removing multiple weights, we revisit one of the classic approaches for impact-based pruning: the Optimal Brain Surgeon(OBS). We propose a tractable heuristic for solving the combinatorial extension of OBS, in which we select weights for simultaneous removal, as well as a systematic update of the remaining weights. Our selection method outperforms other methods under high sparsity, and the weight update is advantageous even when combined with the other methods.
Speaker: Thiago Serra, Ph.D.
Assistant Professor, Bucknell University
Thiago Serra is an Assistant Professor of Analytics and Operations Management at Bucknell University's Freeman College of Management and Director of the Advanced Analytics Research Lab at Bucknell. Previously, he was a Visiting Research Scientist at Mitsubishi Electric Research Labs from 2018 to 2019, and an Operations Research Analyst at Petrobras from 2009 to 2013. He has a Ph.D. in Operations Research from Carnegie Mellon University's Tepper School of Business, and received the Gerald L. Thompson Doctoral Dissertation Award in Management Science in 2018. Dr. Serra was also awarded the INFORMS Judith Liebman Award in 2016, the Best Poster Award at the Princeton Day of Optimization in 2018, and the AAAI Outstanding Program Committee Award in 2021. Dr. Serra‚Äôs research at the intersection of artificial intelligence and mathematical optimization has been funded by the National Science Foundation and published in conferences such as ICML, CPAIOR, AAAI, and NeurIPS as well as in journals such as Mathematical Programming, INFORMS Journal on Computing, and Annals of Operations Research. Dr. Serra currently serves as the Vice Chair / Chair-Elect of the INFORMS Computing Society.
Leveraging Social Media for Urban Analytics and Social Perception
Gathering and analyzing the perception of users towards a service can aid in constructing a more comprehensive service evaluation, one that accounts for community needs and can properly address qualitative aspects of city services. This is only one way to leverage social media and social perception; from urban examples to disaster management and social movements, social media can help us generate insights for decision making and data-driven analysis.
Speaker: Gabriela Gongora-Svartzman, Ph.D.
Assistant Teaching Professor, Heinz College of Information Systems and Public Policy, Carnegie Mellon University
Gabriela Gongora-Svartzman is an Assistant Teaching Professor for Heinz College of Information Systems and Public Policy, at Carnegie Mellon University. Gabriela holds a Ph.D. in Engineering Management from Stevens Institute of Technology, School of Systems and Enterprises. She also has a double major as a B.Sc. in Computer Engineering and Electrical Engineering and an M.Sc. in Computer Science. Her research interests are in data visualization for decision making, urban analytics, social perception, and resilience to disruptions in city services and natural disasters. Gabriela has served as a Session Chair for the PSOR cluster (2018-2021), where she has organized sessions such as the "Diversity/PSOR/MIF - Diversity, Equity, and Inclusion in OR/MS/Analytics. Innovations in Research and Practice" and "Resilient Infrastructure and Community Networks," to highlight a couple. She has also served as Co-chair of the Early Career Teachers Network (ECTN 2021-2022), Secretary for Women in OR/MS (2021), current VP of Communications for Women in OR/MS (2022) and has been both mentor and mentee for the WORMS Mentorship program (2018-2021).