Giving AI a “crystal clear” view of the materials world

MIT researchers developed a multimodal AI model that integrates 3D atomic crystal structures with natural language descriptions to better predict material properties and applications. This approach outperforms text-only models by enabling the AI to understand atomic arrangements, aiding in discovering materials like solid-state lithium-ion battery electrolytes. The model bridges the gap between atomic structure and material performance, advancing materials science AI capabilities.