{"version":1,"type":"story","url":"https://digestai.news/story/mit-researchers-unveil-multimodal-ai-model-that-links-crystal-structur","json":"https://digestai.news/story/mit-researchers-unveil-multimodal-ai-model-that-links-crystal-structur.json","markdown":"https://digestai.news/story/mit-researchers-unveil-multimodal-ai-model-that-links-crystal-structur.md","slug":"mit-researchers-unveil-multimodal-ai-model-that-links-crystal-structur","headline":"MIT researchers unveil multimodal AI model that links crystal structures to properties","summary":"MIT scientists have built a vision‑language model that fuses three‑dimensional atomic structures with natural‑language descriptions of crystals. The team, led by Ju Li and including Yang Jeong Park, Mayank Kumaran, Chia‑Wei Hsu, and Elsa Olivetti, trained the system with contrastive learning on AI‑generated narratives and computational databases, creating unified textual summaries of composition, structure, properties, processing, and applications.\n\nIn a paper published in *Nature Communications*, the researchers compared a version of the model trained on the multimodal dataset with a baseline trained only on text from papers. The multimodal model identified better material designs, correctly surfacing candidates already known for solid‑state lithium‑ion battery electrolytes, and learned geometric representations that text‑only models missed. Li likened the approach to AlphaFold’s linking of protein structure to function, suggesting it could become a step change for materials science.\n\nThe work was supported in part by the U.S. Defense Threat Reduction Agency, the U.S. Defense Advanced Research Projects Agency, and the National Research Foundation of Korea, highlighting cross‑agency interest in accelerating materials discovery.","keyPoints":["MIT team created a vision‑language model that combines 3D crystal structures with textual descriptions.","The multimodal model outperformed text‑only baselines, identifying known solid‑state battery electrolyte candidates.","Training used contrastive learning on AI‑generated narratives and computational databases, funded by DARPA and other agencies."],"whyItMatters":"Linking crystal geometry with language lets AI propose new materials, speeding discovery for batteries and other technologies and expanding AI beyond text‑only domains.","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"entities":{"companies":["MIT","U.S. Defense Threat Reduction Agency","U.S. Defense Advanced Research Projects Agency","National Research Foundation of Korea"],"models":[],"people":["Ju Li","Yang Jeong Park","Mayank Kumaran","Chia-Wei Hsu","Elsa Olivetti"]},"firstPublishedAt":"2026-10-18T04:00:00Z","updatedAt":"2026-10-18T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"MIT News on AI","title":"Giving AI a “crystal clear” view of the materials world","url":"https://news.mit.edu/2026/giving-ai-crystal-clear-view-of-materials-world-1018","publishedAt":"2026-10-18T04:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"MIT researchers unveil multimodal AI model that links crystal structures to properties\", 18 October 2026, https://digestai.news/story/mit-researchers-unveil-multimodal-ai-model-that-links-crystal-structur","publisher":"Digest AI","title":"MIT researchers unveil multimodal AI model that links crystal structures to properties","datePublished":"2026-10-18T04:00:00Z","url":"https://digestai.news/story/mit-researchers-unveil-multimodal-ai-model-that-links-crystal-structur"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}