A3. Information Geometry

Special Session organized by: D. Johnston, H. Matsuzoe, G. Ruppeiner and T. Wada

Information geometry: its connection to statistical mechanics and related fields

Representing probability geometrically in terms of a Riemannian metric is an idea dating back to Fisher and Rao. That the resulting stochastic manifolds have a use in statistical mechanics and thermodynamics has been demonstrated more recently by a number of researchers.
The goal of this workshop is to communicate the message that information geometry has something to say about statistical mechanics amounting to more than just a repackaging of standard textbook topics. Although all applications of information geometry to statistical mechanics will be considered for this session, especially welcome are results from model evaluations, quantum phase transitions, geometry of thermodynamics, stochastic thermodynamics, thermodynamic curvature, and black hole thermodynamics. Also especially welcome would be contributions in the areas of statistics, machine learning, control theory, computer vision, optimal transport theory, and Wasserstein geometry.