ALOP Colloquium with Prof. Andrés Gómez

Assistant Professor Andrés Gómez of Daniel J. Epstein Department of Industrial and Systems Engineering, University of Southern California, will speak at our colloquium on June 15, 2020 at 5 pm about his research via ZOOM.

Title: Sparse learning via mixed-integer optimization.

Abstract: This talk addresses inference problems in which sparsity plays a fundamental role. Sparse statistical models are interpretable by design and often have improved out-of-sample performance than dense models, but learning a sparse model model from data often results in a combinatorial optimization problems. Unfortunately, off-the-shelf mixed-integer optimization software struggle with the resulting problems, resulting in prohibitive solution times even in small instances. We discuss techniques to improve the performance, often related to the convexification of the underlying combinatorial problems.

For more details regarding the Zoom link, please refer back to this page.