DC optimization for Machine Learning and Data Science

• Grant number: 24/21644-8
• Support Opportunities: Scholarships in Brazil - Doctorate (Direct)
• Start date: March 01, 2025
• End date: July 31, 2029
• Field of knowledge: Physical Sciences and Mathematics - Mathematics - Applied Mathematics
• Principal Investigator: Paulo José da Silva e Silva
• Grantee: Giovana Melo dos Santos
• Host Institution: Instituto de Matemática, Estatística e Computação Científica (IMECC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated research grant: 23/08706-1 - Numerical Optimization, AP.TEM

Abstract

This research project focuses on Difference of Convex (DC) optimization applied to Machine Learning and Data Science. DC optimization, where the objective function is the difference of two convex functions, allows the use of convex analysis techniques and offers a flexible modeling framework. The project aims to investigate specific Machine Learning models, such as Support Vector Machines, which appear well-suited for DC optimization. By leveraging DC techniques, we aim to improve generalization performance and robustness while developing more efficient algorithms. A promising direction is incorporating inertia information, as explored in recent research, to implicitly include second-order information within the optimization process, which seems particularly suitable for DC problems.