Our Research
Led by Cornell University, NSF AI-MI is accelerating the discovery of next-generation materials essential for energy, sustainability and quantum technologies. By bringing together computer scientists, materials researchers and data scientists, AI-M tackles knowledge- and data-centric challenges to advance AI and materials science.
Eun-Ah Kim presents research on machine learning. Photo credit: Dave Burbank
Hard Material
Theme Lead: Eun-Ah Kim
The theme encompasses four focus areas: superconductivity, moiré materials, bulk synthesis, and film synthesis. With the superconductivity and moiré materials, the main objective is to transition from a serendipitous mode of discovery to an insight- or prediction-driven discovery. We aim to build AI models that can predict superconductivity or desirable moire settings based on training data. With bulk and film synthesis, the primary objective is to accelerate and optimize synthesis by leveraging accumulated synthesis data.
Soft Material
Theme Lead: Peter Frazier
The Soft Materials Theme develops and applies AI methods to accelerate the design and discovery of advanced soft materials in grand challenge areas. The theme targets problems with significant scientific impact and societal relevance while demonstrating AI’s transformative potential in materials science.
Peter Frazier discusses artificial intelligence approaches with a collegue. Photo credit: Sreang Hok
Students review Jennifer Sun’s research on LLM optimization.
Science-Ready Large Language Model (Sci-LLM)
Theme Lead: Jennifer Sun
Sci-LLM is a specialized AI system designed to help materials scientists overcome challenges related to fragmented knowledge and complex, diverse datasets. It provides a natural-language interface that enables researchers to retrieve information from scientific literature and experimental data, generate actionable code, and interact with AI algorithms, shifting from intuition-driven discovery to AI-augmented materials design.
AI Materials Science Ecosystem (AIMS-EC)
Theme Lead: Anil Damle
We are creating the AI Materials Science Ecosystem, an open, cloud-based portal that integrates a science-ready large language model with data from experiments, simulations, images and scientific literature. Through the AIMS-EC, the institute will harness AI to unlock next-generation discoveries such as resilient qubits for quantum computing, new superconductors for faster and cheaper electronics, peptides that remove microplastics from the environment and advanced materials for sustainable manufacturing.
Anil Damle presents his research on quantum localization methods. Photo credit: IPAM at UCLA
Publications
Explore AI-MI research publications advancing the discovery of next-generation materials. These works reflect interdisciplinary collaboration across AI, materials science, and data science, showcasing methods, insights, and results that accelerate innovation and enable reproducible, scalable research.