- Artificial Intelligence
- Data-Driven Approaches
- Engineers
- Machine Learning
- Research Network and Facility Services Division Platforms
Pioneering Material DX Conquering the Challenging Path to Success
Pioneering Material DX Conquering the Challenging Path to Success
The Ministry of Education, Culture, Sports, Science and Technology (MEXT) is promoting the development of the “Material DX Platform” through the integrated efforts of three interlinked initiatives: the Advanced Research Infrastructure for Materials and Nanotechnology in Japan (ARIM), the Materials Data Platform (MDPF), and the Data-creation and Application-oriented Materials Research and Development Project (DxMT). Among these, NIMS serves as the sole implementing organization for the MDPF initiative, responsible for collecting and accumulating materials data from across Japan and developing the necessary infrastructure and frameworks for its utilization. In this interview, Satoshi Minamoto, Takuya Kadohira, and Isao Kuwajima, who oversee the MDPF initiative, share insights into its current status and future prospects.
Satoshi Minamoto
Materials Data Platform, Platform Director / Data Application Unit, Unit Leader
Takuya Kadohira
Materials Data Platform, Deputy Platform Director / Data Infrastructure Unit, Unit Leader
Isao Kuwashima
Materials Data Platform, Data Collection Unit, Unit Leader
The Core of the Data Strategy: The Role of MDPF
—Could you tell us about the MDPF, one of the key pillars of the Material DX Platform development?
Minamoto: At MDPF, our goal is to collect, accumulate, distribute, and utilize high-quality materials data. We are expanding various data services while developing the necessary infrastructure. Our concept is to seamlessly manage the entire process—from data input (creation) to accumulation (collection) and output (utilization).
Kadohira: Since the mid-2010s, NIMS has been implementing an internal cycle of “creation, collection, and utilization” of materials data. Entering the 2020s, under the Material DX Platform initiative, we have expanded this cycle to form a broader network that includes research institutions and universities across Japan. We are currently focused on developing and implementing the necessary steps to realize this vision.
Kuwajima: Within MDPF, I lead the Data Collection Unit, which is responsible for gathering materials data. Dr. Kadohira manages the Data Infrastructure Unit, overseeing the infrastructure development, while Dr. Minamoto leads the Data Application Unit, promoting the practical use of material data. However, in practice, our units aren’t strictly divided. We all share the common goal of advancing MDPF through cutting-edge innovation while ensuring stable operations. We work together as a unified team towards this objective.
—The first step in the data strategy is to collect and accumulate materials data. How is this progressing?
Kuwajima: Since 2003, NIMS has been developing MatNavi, one of the world’s largest materials databases, comprised of over a dozen databases and several applications covering inorganic materials, metallic materials, polymers, and more. The Data Collection Unit extracts relevant data from research papers, integrates experimental results from NIMS, and performs calculations when necessary. The data are then standardized and registered in the databases. Expert curation ensures the accuracy of information from papers, while materials testing—such as creep testing—are conducted according to consistent standards at NIMS. This maintains MatNavi’s exceptional data quality, making it one of the core services of MDPF.
Kadohira: Alongside the development of MatNavi, MDPF has also been enhancing data collection through a service called RDE (Research Data Express). RDE is a system designed to transfer data directly from remote experimental instruments and store it in a format optimized for AI-based analysis. Conventionally, data output from experimental instruments varied in terminology and format, making it difficult to integrate into a unified database. However, RDE features an automated structuring function that translates terminology, assigns metadata, and converts formats. This enables real-time data archiving in MDPF’s shared database.
Initially, RDE was operated within a closed environment at NIMS to establish an internal data service cycle. In January 2023, RDE was officially incorporated into the MDPF service lineup, and MDPF’s system infrastructure was migrated to a public cloud. This transition enabled experimental instruments from research institutions and universities participating in ARIM to transmit data online via RDE, centralizing data collection and storage. This integration established the full-scale collaboration of ARIM, MDPF, and DxMT into an interconnected system.
