Research

Advancing interdisciplinary research at the intersection of materials science, data science, and emerging technologies.

📊

Materials Informatics & AI

Materials Informatics (MI) is revolutionizing how we discover and design new materials by integrating computational methods, machine learning, and experimental data. Our research focuses on developing AI-driven approaches to accelerate materials discovery and optimize properties for various applications.

Key Research Topics

  • ▸ GPU-accelerated Analysis Systems: Development of high-performance computing tools for large-scale materials data analysis, enabling rapid screening of materials candidates.
  • ▸ Machine Learning Models: Building predictive models for materials properties using advanced ML techniques including graph neural networks and transfer learning.
  • ▸ Thermoelectric Materials Mapping: Comprehensive mapping of thermoelectric materials using integrated computational and experimental datasets.
  • ▸ Industry-Academia Collaboration: Bridging academic research with industrial applications through collaborative projects with manufacturing companies.
🤖

Research DX & IoT Innovation

Digital Transformation (DX) in research environments is essential for accelerating scientific discovery. We develop and implement IoT systems, automation tools, and AI-powered platforms to enhance research productivity and enable new forms of interdisciplinary collaboration.

Key Research Topics

  • ▸ IoT Systems Development: Deployed at 8 Tohoku University locations for environmental monitoring and automated data collection.
  • ▸ Natural Language Processing: Development of NLP-based matching systems to facilitate researcher collaboration and knowledge discovery.
  • ▸ Research Automation: Creating automated pipelines for data acquisition, processing, and analysis in experimental settings.
🔬

Ultrafast Magneto-Optical Spectroscopy & Spintronics

Understanding the ultrafast dynamics of magnetic materials is crucial for developing next-generation spintronic devices. Our research employs advanced time-resolved magneto-optical imaging (TRMOI) techniques to study spin dynamics at unprecedented temporal and spatial resolutions.

Key Research Topics

  • ▸ Time-Resolved Magneto-Optical Imaging: Achieving 10^-12 second time resolution for observing ultrafast magnetization dynamics.
  • ▸ Spin Wave Tomography: All-optical techniques for mapping spin wave dispersion in magnetic materials with 10^-6 meter spatial resolution.
  • ▸ Photo-Induced Magnetization: Investigating light-induced magnetic phenomena for potential applications in ultrafast data storage and processing.
  • ▸ JST ERATO Project: Core technology development for JST ERATO project on ultrafast magnetization control.
⚗️

Autonomous Experiment Systems

We are developing a self-driving chemical synthesis system for functional materials (MOFs, nanoparticles, etc.) using robotic arms and electric pipettes, which has reached the proof-of-concept stage. The system features a closed-loop architecture that links AI-driven experiment condition optimization with robotic automated execution.

Key Research Topics

  • ▸ Closed-Loop AI Optimization: Bayesian optimization algorithms autonomously determine the next synthesis conditions, the robot executes the experiment, and the model is updated based on the results.
  • ▸ High-Precision Robotic Control: Python-based integrated control for precise reagent dispensing at the microlitre level, image recognition-based position correction, and reaction temperature feedback control.
  • ▸ Autonomous vs. Automated: Moving beyond conventional automation (repeating fixed procedures) to truly autonomous exploration that efficiently navigates the experimental parameter space.

Current Projects

Nano-Materials Process Data Science

Endowed research division focusing on AI-driven materials research. Integrating computational and experimental approaches for advanced materials discovery.

Endowed Division • Tohoku University

AI Terakoya Educational Platform

Comprehensive video learning platform with 30+ series and 140+ chapters covering Materials Informatics, Materials Science, and Machine Learning.

Education • Open Access