Masaya Taniguchi

Biography

Masaya Taniguchi is a JSPS Postdoctoral Research Fellow in the Explainable AI Team at the RIKEN Center for Advanced Intelligence Project (AIP). He studies how the order and history of learning shape the structures and capabilities acquired by computational systems. His research combines mathematical analysis and controlled experiments on learning dynamics, curriculum learning, and data-order control, building on categorial grammar, substructural logic, and formal-language theory. He also works on verifiable AI and AI for Mathematics, connecting language models with theorem proving, formal verification, and countermodel generation. He earned his B.S. in Mathematics from Gakushuin University and his M.S. and Ph.D. in Information Science from JAIST.

Research interests

• Learning Dynamics and Training-History Dependence
• Curriculum Learning and Data-Order Control
• Learnability and Expressivity of Language Models
• Neuro-Symbolic AI and AI for Mathematics
• Automated Theorem Proving and Formal Verification
• Natural Language Processing
• Formal Semantics and Grammar of Natural Language
• Combinatory Categorial Grammar
• Substructural Logic and Proof Theory
• Cyclic Proof Systems and Lean Formalization
• Programming Languages and Proof Assistants