AI can predict the physical properties of a material just by 'reading' its chemical name, no 3D modeling required.
April 14, 2026
Original Paper
ReadMOF: Structure-Free Semantic Embeddings from Systematic MOF Nomenclature for Machine Learning
arXiv · 2604.10568
The Takeaway
By using systematic nomenclature, models can predict properties of Metal-Organic Frameworks with accuracy comparable to full 3D atomic simulations. It suggests that chemical language contains enough structural data that a model can 'read' properties without seeing physical shapes.
From the abstract
Systematic chemical names, such as IUPAC-style nomenclature for metal-organic frameworks (MOFs), contain rich structural and compositional information in a standardized textual format. Here we introduce ReadMOF, which is, to our knowledge, the first nomenclature-free machine learning framework that leverages these names to model structure-property relationships without requiring atomic coordinates or connectivity graphs. By employing pretrained language models, ReadMOF converts systematic MOF na