The birth of information geometry dates back to the mid-20th century. Starting with the discovery of dual connections inherent in families of probability distributions, it has led to a wide range of applications in science.
In light of the advancements in data science today, the significance of information geometry has once again been acknowledged. The purpose of this international conference is to facilitate the exchange of recent developments in information geometry and to establish its theoretical foundations across related fields. The development of information geometry will be explored through a series of invited lectures by leading researchers working in mathematics, machine learning, statistics and general natural sciences, as well as short talks and poster presentations.
Nihat Ay is giving an invited lecture on “Information Geometry of the Otto Metric”, and Adwait Datar will contribute a short talk on "Convergence Properties of Natural Gradient Descent for Minimizing KL Divergence”.
The research meeting will be held from March 18 (Tue) to 21 (Fri), 2025, at Hongo Campus, The University of Tokyo, in Tokyo, Japan. There will be a tutorial course in Japanese on March 17 (Mon).
More information
