3S-G09
Data-driven R&D zone
Materials R&D in the AI era: physics, simulation and data
We will showcase three complementary technologies for materials R&D: the simulation-integrated materials informatics platform “MI-Suite,” the multiscale materials simulation platform “J-OCTA,” and “ASAP/ASAP-PRO” for first-principles calculations and machine-learning force fields. Together, they connect electronic and atomic-scale physics, molecular and mesoscale simulation, and experimental and simulation data analysis to support materials development in the age of AI.
MI-Suite is a materials informatics platform that integrates experimental and simulation data for materials R&D. It supports data management, preprocessing, analysis, machine learning, property prediction and materials exploration. Physics-based information obtained from simulations, including molecular interactions, interfaces, free volume, structural features and dynamics, can also be incorporated into the analysis. MI-Suite is designed to connect data-driven prediction with physical information that can support the next decision in materials design.
J-OCTA is a multiscale simulation platform for studying materials phenomena from atomic and molecular scales to the mesoscale. It provides modeling and analysis capabilities for molecular interactions, interfaces, diffusion, phase separation, dispersion, viscoelasticity and other structure–property relationships. Building on more than two decades of research and development originating from the OCTA project, J-OCTA is also evolving as a physics-based simulation platform that can be combined with AI and materials informatics.
ASAP / ASAP-PRO provides atomic-scale materials simulation based on first-principles calculations for studying electronic structures, defects, interfaces, reactions and diffusion. Machine-learning force fields can extend atomistic simulations beyond the size and time scales that are practical with DFT alone. ASAP can also be combined with J-OCTA and MI-Suite to connect electronic- and atomic-scale information with multiscale simulation and materials data analysis. ASAP-PRO also supports workflows based on MACE machine-learning force fields, with DFT reference calculations and validation used to assess applicability to target materials.
-
2026/12/18 13:00 ~ 13:30Venue:Seeds & Needs Seminar A(South hall 3)
Select this
session【Abstract】Rapid advances in artificial intelligence (AI) and machine learning are transforming materials development through data-driven prediction, materials exploration, and accelerated simulation. However, the properties of real materials are affected not only by their chemical composition but also by hierarchical structures spanning atomic, molecular, mesoscopic, and macroscopic scales, together with processing history. Physics-based simulation therefore continues to play an important role in materials design. This presentation reviews research on multiscale modeling accumulated through OCTA/J-OCTA and discusses its recent integration with AI, including structure recognition, surrogate modeling, and optimization. We also introduce MI-Suite, which aims to integrate experimental and simulation data for analysis and materials exploration. Through these examples, we discuss how multiscale simulation can evolve and complement AI in future materials development.
Dr. Taku Ozawa
JSOL CORPORATIONEngineering Technology Business UnitGeneral Manager【Profile】Taku Ozawa is engaged in the development and commercialization of materials simulation software at JSOL Corporation, including OCTA/J-OCTA. His interests include multiscale simulation from molecular to mesoscopic and continuum scales, materials informatics, and the integration of simulation with AI. He holds a Ph.D. and serves on a computational science research group within the Society of Polymer Science, Japan.
*The following information is provided to enable inquiries to exhibitors.
Unauthorized use and reproduction for any other purpose is prohibited.
Engineering Technology Business Unit
Apply for seminar