David Alvarez-Melis

Kempner Institute Associate Faculty
Assistant Professor of Computer Science

Contact Information


David Alvarez-Melis, Assistant Professor of Computer Science, researches fundamental principles of artificial intelligence, with a particular focus in understanding, characterizing, and advancing methods for learning in constrained (e.g., data-scarce), dynamic (e.g., time-drifting), and complex (e.g., multi-modal) environments—all settings where natural intelligence excells. Alvarez-Melis’s work takes a data-centric approach, focusing on studying and intervening the inputs and latent representations of deep learning models using statistical-geometric methods like optimal transport theory. Much of his recent research is motivated by problems in the natural sciences and seeks to develop “scientific artificial intelligence” by finding a deeper understanding —and in-silico emulation— of various aspects of human intelligence that are fundamental for scientific thinking and discovery.