Wotao Yin obtained his Ph.D. in 2006 under the supervision of Prof. Donald Goldfarb. He is a scientist of Alibaba US DAMO Academy and director of its Decision Intelligence Lab.
Title of talk:
Solving LPs, QPs, MILPs, and MIQPs Using Graph Neural Networks
This presentation explores the relationship between Graph Neural Networks (GNNs) and mathematical optimization, including Linear Programming (LP), Quadratic Programming (QP), Mixed-Integer Linear Programming (MILP), and Mixed-Integer Quadratic Programming (MIQP). By defining optimization instances on specific graph structures, GNNs can provably assess problem feasibility, unboundedness, and compute solutions to any desired precision, within limits on the instances' symmetry properties. These results not only enhance our understanding of GNNs' expressive capabilities but also open new opportunities for leveraging these models to accelerate numerical optimization processes.
