image: SMU engineering professor Digvijay Boob won a five-year CAREER grant from the National Science Foundation to come up with quicker, streamlined solutions that could speed up how AI learns from data to make predictions and decisions.
Credit: SMU
DALLAS (SMU) – The National Science Foundation (NSF) has awarded SMU engineering professor Digvijay Boob a five-year CAREER grant to pioneer quicker, streamlined solutions that could speed up how AI learns from data to make predictions and decisions.
The applications cover a wide range of tasks, such as ensuring fairness in machine learning models, improving inventory management for healthier supply chains and making optimal decisions in controlling power plants.
Boob will receive a $596,607 grant from the CAREER Program, NSF’s most prestigious award in support of early-career faculty. His project is aimed at addressing algorithms for two types of mathematical problems. Known as semi-infinite optimization problems and equilibrium-constrained optimization problems, these math problems are used by scientists and business practitioners to improve efficiency, reduce costs, manage risks, enhance decision-making, and maximize profits. Data scientists and engineers also use these optimization problems in their field to assess predicted outcomes and make smart decisions.
“We do not currently have sufficient understanding of how to solve these problems,” said Boob, an assistant professor in Operations Research and Engineering Management at SMU’s Lyle School of Engineering. “Existing algorithms tend to only work in limited settings, or else they can’t handle the large-scale problems we are facing in today’s world.”
Machine learning, a subset of artificial intelligence (AI), empowers computers to learn from data, enabling them to make predictions or decisions without explicit programming. Algorithms are the backbone of machine learning models, providing the mathematical instructions guiding AI’s responses to the data it encounters.
Faster algorithms, such as those Boob aims to develop, could potentially solve optimization problems more quickly than current methods. This is crucial as AI models become more complex to meet the expanding demands of users.
Semi-infinite optimization problems are mathematical problems with an infinite number of potential constraints. Picture a plant that manufactures or processes chemicals; to avoid unintended chemical reactions, several variables must be kept in mind and not all of them can be anticipated in advance. A semi-infinite optimization problem can be used to ensure that all of those constraints are taken into account to get the desirable outcome.
“Equilibrium-constrained optimization problems,” on the other hand, “are used to model a system where multiple players act independently in self-interest,” Boob explained. An example of this would be competing companies using the algorithm to figure out how they can maximize their profits while considering the reactions of competitors.
“The job of the algorithm is to find good equilibriums despite uncertain aspects,” said Boob, who specializes in developing provably fast-converging, easy-to-implement scalable algorithms.
However, writing algorithms for these different optimization scenarios will be extremely difficult.
“Not only do the solution points need to satisfy the constraints, but they have to be the best among all such points, either maximizing some profit or minimizing errors while respecting those infinite or equilibrium constraints,” Boob said. “Clearly, a haphazard approach will only take us so far. We want to study these problems systematically.”
SMU students from the Office of Engaged Learning who have a strong background in mathematics, optimization, and computer language will be helping Boob test how well the algorithms work and implement them on a computer to see them running in action.
This material is based on work supported by the National Science Foundation under Grant No. 2340858. The NSF’s CAREER Program is a Foundation-wide activity that offers the National Science Foundation's most prestigious awards in support of early-career faculty who have the potential to serve as academic role models in research and education and to lead advances in the mission of their department or organization. It is expected that activities pursued by early-career faculty, as part of this award, will build a firm foundation for a lifetime of leadership in integrating education and research.
Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.
About SMU
SMU is the nationally ranked global research university in the dynamic city of Dallas. SMU’s alumni, faculty and more than 12,000 students in eight degree-granting schools demonstrate an entrepreneurial spirit as they lead change in their professions, communities and the world.