Vishal Gupta
Dean’s Associate Professor of Data Sciences and Operations
USC Marshall School of Business
Associate Professor (by courtesy) of Industrial & Systems Engineering
USC Viterbi School of Engineering
- Associate Editor · Management Science, MSOM, Operations Research
- Co-PI (USC site) · AI4OPT, the NSF AI Institute for Advances in Optimization
- Affiliate · ORAI PhD Certificate, Kendrick Supply Chain Institute, Center for AI in Society
I'm recruiting 2–3 PhD students for Fall 2027.PhD Students →
Research
I study how the structure of optimization problems should shape learning, especially when data are scarce or noisy. These settings are surprisingly common. AI applications draw on huge datasets to make thousands of personalized, real-time decisions, but the data relevant to any one decision are often limited.
My work develops methods that use the decision problem’s own structure as an inductive bias. I create methods that exploit that structure, study the theory of what we can and can’t learn, and, where possible, design systems that are deployed.
Optimization structure as a design principle for learners
How should a decision problem’s structure shape what we learn? Often, that structure dictates the right learning loss and the right architecture.
More papers in this stream →Small-Data, Large-Scale
How do common methods for large-scale decision-making behave when data are scarce? I develop theory to answer that, and methods that do better.
More papers in this stream →Data Scarcity in Practice
Real-world data are often aggregated, redacted for privacy, or generated by logged policies with little variation. How do we design systems from such data?
More papers in this stream →Selected Publications
- Decision-Focused Learning with Directional Gradients
with M. Huang* · NeurIPS, 2024
- Debiasing In-Sample Policy Performance for Small-Data, Large-Scale Optimization
with M. Huang* and P. Rusmevichientong · Operations Research, 2024
- Data-Pooling in Stochastic Optimization
with N. Kallus · Management Science, 2022
- Efficient and Targeted COVID-19 Border Testing via Reinforcement Learning
H. Bastani†, K. Drakopoulos†, V. Gupta†, J. Vlachogiannis, C. Hadjichristodoulou, P. Lagiou, G. Magiorkinis, D. Paraskevis and S. Tsiodras · Nature, 2021
† Equal contribution. Author order is not alphabetical.
Spotlight Presentation at the Reinforcement Learning for Real-Life Workshop (ICML 2021)
Winner, 2021 Pierskalla Best PaperWinner, 2021 Wagner Prize (for the Eva system)
Programs I Lead
DSO Summer Scholars
I founded and co-direct DSO Summer Scholars, which brings USC undergraduates into a summer of real research with faculty.
Visit the program →BUAI Builder Hub
I’m the inaugural faculty advisor for the BUAI Builder Hub, where students work with faculty to build AI tools for the university and the Los Angeles community.
Visit the hub →