The arrow of time: at the intersection of thermodynamics, machine learning, and causality
By Alberto Suárez
Alberto Suárez will give a talk on the arrow of time: at the intersection of thermodynamics, machine learning, and causality.
Abstract
The arrow of time refers to the asymmetry in the evolution of physical systems. It is characterized by the second law of thermodynamics. This statistical law states that, in an isolated system, entropy cannot decrease with time and is constant if and only if all processes are reversible. Since microscopic dynamics are reversible, time’s arrow is an emergent property that is apparent only at the meso- and macroscopic levels, both of which involve loss of detail. Machine learning by automatic induction is also an asymmetric dynamical process in which the identification of patterns involves some information loss. The asymmetry of time plays a role also in causal inference: causes precede effects. Finally, causal explanations, which are ubiquitous in human reasoning, are key to rendering machine learning models interpretable. In this talk we will review recent work around these ideas to uncover relations between thermodynamics, machine learning, and causal inference that could provide fruitful insights into the emergence of meaning from raw data.
Biography
Alberto Suárez received the degree of Licenciado (BSc) in Chemistry, specialization in Quantum Chemistry, from the Universidad Autónoma de Madrid, Spain, in 1988, and the PhD in Physical Chemistry from the Massachusetts Institute of Technology (MIT), Cambridge, MA, in 1993. After holding postdoctoral positions at Stanford University (USA), at Université Libre de Bruxelles (Belgium), as a research fellow financed by the European Commission within the Marie Curie “Training and Mobility of Researchers” program, and at the Katholieke Universiteit Leuven (Belgium), he is currently Professor of Computer Science and Artificial Intelligence in the Computer Science Department at the Universidad Autónoma de Madrid (UAM), where he co-directs the Machine Learning Group - Grupo de Aprendizaje Automático (MLG-GAA) [www.eps.uam.es/~gaa]. He has also held appointments as “Senior Visiting Scientist” at the International Computer Science Institute (Berkeley, CA) and at MIT (Cambridge, MA). He has worked on relaxation theory in condensed media, stochastic and thermodynamic theories of nonequilibrium systems, lattice-gas automata, and automatic induction from data. His current research interests include artificial intelligence, in particular, machine learning, computational statistics, functional data analysis, and causality. He is a member of IEEE, of the European Laboratory for Learning and Intelligent Systems (ELLIS) and a founding member of ELLIS Unit Madrid.–