IIT Madras Builds AI Platform Holding 185,000 Alloy Records to Speed Materials Discovery

Researchers at IIT Madras have built an artificial intelligence platform intended to speed up the discovery of sustainable, high-performance metallic alloys for electric vehicles, aerospace, renewable energy and marine infrastructure.

The team used large language models to extract and organise information from more than 10,000 research papers, producing two databases holding over 185,000 structured records. IIT Madras describes them as the world's largest publicly available multicomponent alloy databases, and they are free to use through the Alloy Tattvasar platform and GitHub.

The work targets a practical obstacle in materials research: useful experimental detail is scattered across journal articles, tables and figures, which makes systematic comparison slow. The automated pipeline extracts alloy compositions, manufacturing processes, testing conditions and more than 350 material properties, keeping the conditions under which each measurement was taken. It uses Retrieval-Augmented Generation to pull relevant examples in while extracting, which the researchers say improves accuracy on text and tables.

"Instead of spending years manually collecting data from thousands of publications, our framework automatically builds structured databases that can be used to identify sustainable materials much faster," said Rohit Batra, assistant professor in the Department of Metallurgical and Materials Engineering.

The platform also records environmental, economic and social indicators alongside performance, so candidates can be judged on sustainability as well as strength. The team demonstrated it by identifying high-entropy alloys for lightweight automotive and aerospace structures, soft magnetic materials for electric motors and transformers, and corrosion-resistant alloys for marine and offshore engineering. The study was published in Advanced Science, and the group plans to extend the framework to figures and microstructural images, and to polymers, ceramics and composites.