Working with databases in the field of materials research currently requires users to have appropriate programming knowledge in many cases. Researchers who specifically search for materials often have to formulate complex queries.
AI in materials research
With the specially developed AI-supported materials database “StableOx-Cat”, researchers at Tohoku University in Japan now want to lower this entry barrier. The database is intended to enable queries in natural language and at the same time provide scientifically reliable results, according to a statement from the university.
Editorial recommendations
${content}
${custom_anzeige-badge}
${custom_tr-badge}
${section}
${title}
Specifically, the system combines a large language model (LLM) with physically based methods. Similar to a conversation with an AI chatbot, users can ask simple questions to identify materials of interest for their research.
Agent translates natural language
According to the researchers, the AI agent integrated into the material database translates these questions into structured scientific analyses. The system should be able to assess whether a material is stable under various conditions, such as changes in acidity or electrical potential. Both critical factors in real applications.
StableOx-Cat should avoid misleading or incorrect results due to AI hallucinations because the analysis is based on known physical principles, according to the materials researchers at Tohoku University. With the results provided, users can then carry out corresponding experiments in the laboratory.
AI-supported database expandable
The system is currently specialized for the identification of metal oxide electrocatalysts. These materials play a key role in processes such as water splitting and fuel production. However, the platform should be expandable to also examine other types of materials such as alloys, nitrides and carbides in the future.
The goal is to discover better materials for the development of clean energy technologies in a shorter time. Combining natural language interaction with rigorous scientific assessment enables more researchers to “explore complex chemical spaces efficiently and safely,” explains materials researcher Hao Li from Tohoku University.
The researchers described the AI system in a study published in the journal AI Agent.
Top Article
${content}
${custom_anzeige-badge}
${custom_tr-badge}
${section}
${title}
