Ryotaro Okabe / 岡部 遼太郎
Ph.D. Candidate in Chemistry, Massachusetts Institute of Technology
Physics-informed AI for quantum materials discovery
I am a Ph.D. candidate in Chemistry at MIT, advised by Prof. Mingda Li in the Quantum Measurement Group. My research develops physics-informed machine learning methods for quantum materials discovery and design, spanning generative models, graph neural networks, and LLM-guided synthesis prediction. Before MIT, I received my M.S. and B.S. from the Institute of Science Tokyo and conducted research internships at IBM Research, Yale, and Rice.
Research Themes
Generative Models for Quantum Materials
Integrating structural constraints into generative AI (SCIGEN) for the inverse design of quantum material candidates.
Graph Neural Networks for Materials Properties
Predicting phonon dispersions, optical spectra, and magnetism directly from crystal structures.
LLMs for Materials Synthesis
Language-model-guided prediction of synthesis routes and conditions for quantum materials.
AI-Assisted Detection and Scientific Instrumentation
Tetris-inspired detector design and neural-network-based radiation mapping.
Selected Publications
Recent Highlights
- 2026-05 Preprint Universal Magnetic Structure Prediction from Atomic Coordinates with Near-Experimental Accuracy posted to arXiv
- 2026 Review Artificial intelligence-driven approaches for materials design and discovery published in Nature Materials
- 2025 Structural Constraint Integration in a Generative Model for the Discovery of Quantum Materials (SCIGEN) published in Nature Materials and featured by MIT News
- 2025-04 Received the MRS Spring Graduate Student Silver Award and Best Poster Award
- 2024-09 Selected for the IBM PhD Fellowship