Machine Learning NeEDS Mathematical Optimization with Prof Leo Liberti

Abstract: We present theoretical and computational results relating to a set of works where we apply random projection techniques to mathematical programming formulation classes, namely linear programs, conic programs, quadratic programs, and one mixed-integer nonlinear program (the k-means problem). Time permitting, we comment on applications of these techniques to the diet problem, quantile regression, compressed sensing, and portfolio optimization.
Joint work with C. D’Ambrosio, B. Manca, P.-L. Poirion, K. Vu.