AI4OPT Seminar Series
Date: Wednesday, April 2nd, 11:30 AM - 1 PM
Location: 9th floor Atrium in Coda Building (756 W Peachtree St NW, Atlanta, GA 30308)
Speaker: Bistra Dilkina
The effective synergy of machine learning and combinatorial solvers
Abstract: We will demonstrate recent advances in successfully integrating machine learning and combinatorial optimization to improve the speed and solution quality of solvers. Many real-world applications that require combinatorial optimization involve solving repeatedly similar instances of the same problem over different data (objective and constraint coefficients). Hence, they provide the opportunity to learn to search more effectively by leveraging historical instances. We design approaches to effectively augment established state-of-the-art Mixed Integer Linear programming (MILP) solvers with ML-guided components to significantly improve performance. In particular, many important (heuristic) tasks in MILP solving involve choosing subsets of variables, and we demonstrate that Contrastive Loss is particularly well-suited for training in this setting by learning from both positive and negative examples of candidate sets. We show the successful application of contrastive loss in the context of Large Neighborhood Search for MILP, as well as Backdoor selection for MILP, resulting in significant speed-ups over multiple domains. We introduce d-MIPLIB (Distributional MIPLIB) as a unified benchmark set for the research community in ML-guided MILP solving in order to facilitate benchmarking across a variety of problem domains and drive progress in this field. We demonstrate the potential of learning across multiple domains when historical examples are limited, and the impressive power of multi-task learning across different ML-guided solving techniques for any given problem domain, setting a key stepping to a foundation model for ML-guided MILP solving.
Bio: Dr. Bistra Dilkina is an associate professor of computer science at the University of Southern California, co-director of the USC Center of AI in Society, and the inaugural Dr. Allen and Charlotte Ginsburg Early Career Chair at the USC Viterbi School of Engineering. She leads the USC Site for the AI Institute for Advances in Optimization (Ai4OPT). Her research and teaching are centered around the integration of machine learning and discrete optimization, with a strong focus on AI applications in computational sustainability and social impact. She received her Ph.D. from Cornell University in 2012 and was a post-doctoral associate at the Institute for Computational Sustainability. Her research has contributed significant advances to machine-learning-guided combinatorial solving, including mathematical programming and planning, as well as to decision-focused learning, where combinatorial reasoning is integrated into machine-learning pipelines. Her applied research in Computational Sustainability spans using AI for wildlife conservation planning, climate change impacts in terms of energy, water, habitat and human migration, and using AI to optimize fortification of lifeline infrastructures for disaster resilience. Her work has been supported by the National Science Foundation, National Institute of Health, DHS Center of Excellence Critical Infrastructure Resilience Institute, Paul G. Allen Family Foundation, Microsoft, and Qualcomm, among others. She has over 90 publications and has co-organized or served as a chair to numerous workshops, tutorials, and special tracks at major conferences.
Note: Lunch will be served at the seminar. So, please stop by 15 minutes before the seminar to pick up lunch.
To continue receiving all AI4OPT seminar announcements, please sign up to the mailing list at https://lists.isye.gatech.edu/mailman/listinfo/ai4opt-seminars
Videos of the past seminars can be seen on AI4OPT webpage at https://www.ai4opt.org/seminars/past-seminars
