MATAI: An AI-Powered Generalist Material Discovery Platform

This project will develop an AI-powered generalist platform with three key components: a holistic material database, a foundational material property predictor, and a generalist multi-property material designer in a large search space.

Lead PI: Bo AN, NTU

Co-PIs:
1. Cuntai GUAN, NTU
2. Zheng LIU, NTU
3. TEO Hang Tong, NTU
4. Bijun TANG, NTU

Collaborators:
1. Carla GOMES, Cornell University
2. Bart SELMAN, Cornell University
3. Pulickel M AJAYAN, Rice University
4. Shufeng KONG, Sun Yat-sen University
5. Yexiang XUE, Purdue University

This project will develop an AI-powered generalist material discovery platform (MATAI), aimed at overcoming the limitations of traditional methods in discovering new materials, particularly high-entropy alloys (HEAs) and multi-functional composites (MFCs). This will comprise three key components: (1) establishing a holistic material database through simulations and experimental feedback, (2) building a foundational material property predictor with a high-dimensional and multi-modal database, and (3) developing a generalist multi-property material designer in a large search space.

The primary goal of developing the MATAI platform is two-fold: (i) identifying novel HEAs and MFCs with enhanced structural and functional properties, and (ii) providing material researchers and AI practitioners an all-in-one solution for efficient material discovery. The platform will be made available to the research community at the end of Stage 1. We will also work closely with academic and industry partners to ensure that the discovered materials can be manufactured at low costs and contribute to the advancement of material research in Singapore.

 

We envision our project to generate the following benefits and potential impacts:

  • For Academia: (i) Material science: MATAI can revolutionize the R&D process of advanced functional materials and generate new knowledge in the domains of materials science, chemistry, and others. (ii) AI: The inverse material design problem presents an ideal testbed for AI algorithms, because of its vast search space, data efficiency requirements, and conflicting multi-property considerations. We believe that MATAI can significantly boost the development of AI algorithms for real-world scenarios.
  • For Industry: The development of novel HEAs and MFCs with enhanced functional properties and lighter weight can be applied in a wide range of industries, from construction and transportation to energy and electronics. This can lead to the creation of new products and services, the generation of new jobs, and eventually accelerate the transition towards Industry 4.0 in Singapore.
  • For Singapore: Our project will complement existing “AI for Materials Discovery” research in Singapore, providing an established first-hand database, cutting-edge algorithms, and nurturing an interdisciplinary team that ensures the sustainable development of related research in the future, placing Singapore at the forefront of this field worldwide.