Singular Learning Theory and Related Topics — At the Intersection of Machine Learning, Mathematical Statistics, and Statistical Physics

Reference No.	2026a029
Type/Category	Grant for Young Researchers and Students-Workshop (II)
Title of Research Project	Singular Learning Theory and Related Topics — At the Intersection of Machine Learning, Mathematical Statistics, and Statistical Physics
Principal Investigator	Naoki Hayashi（Toyota Central R&D Labs., Inc.・Researcher）
Research Period	November 20,2026. - November 20,2026.
Keyword(s) of Research Fields	Bayesian Statistics, Machine Learning, Algebraic Geometry, Algebraic Statistics, and Statistical Physics
Abstract for Research Report	Objective In recent years, research related to singular learning theory has expanded across diverse fields, including machine learning, mathematical statistics, and statistical physics. However, the communities to which researchers belong, as well as their primary venues for dissemination (such as conferences and journals), have become increasingly fragmented. As a result, it is becoming difficult to maintain a comprehensive, cross-disciplinary understanding of developments in these areas. To address this issue, we propose to establish a research workshop under the theme of “Singular Learning Theory” as a platform for interdisciplinary exchange. The goal is to foster continuous interaction and to build a sustainable research network. Expected Outcomes (1) Networking among researchers The workshop will serve as a hub for researchers from academia and industry with diverse backgrounds to gather and exchange information on the latest research developments. It will also create opportunities for collaboration between speakers and participants, fostering an environment conducive to joint research. Through increased mutual recognition and citation among researchers, the initiative is expected to contribute to the maturation of the singular learning theory community. In addition to its connections with machine learning, mathematical statistics, and statistical physics, singular learning theory is closely related to fields such as algebraic geometry and algebraic statistics. The workshop is therefore expected to promote active interaction with these communities as well. Moreover, given the significant presence of industry researchers in this field and the growing interest in applications such as AI alignment, the workshop will also strengthen ties with industry. (2) Development of interdisciplinary talent with foundations in mathematics and statistics The workshop will include tutorial lectures on singular learning theory and its related areas in mathematics (e.g., singularity theory, algebraic geometry, distributions, and empirical processes) and statistics (e.g., Bayesian statistics and statistical machine learning). Through these activities, we aim to cultivate researchers with a hybrid skill set spanning both mathematical and statistical foundations.
Organizing Committee Members (Workshop) Participants (Short-term Joint Usage)	Akifumi Okuno（The Institute of Statistical Mathematics・Assistant Professor） Satoru Tokuda（Kyushu University・Associate Professor）