AI-MI researchers have introduced Graphlet-MP, a database that encodes inorganic crystal structures as graphlet histograms — a data-efficient, interpretable, machine-learning-ready representation aimed at accelerating materials discovery. Covering 149,082 inorganic crystals from the Materials Project, it describes each material by 79 distributions across three hierarchical graphlet orders: atomic sites, bonded pairs, and bond-angle triplets. The release includes a full technical specification, an Earth Mover’s Distance metric for comparing materials, the precomputed database, and open-source code to generate representations for arbitrary crystal structures. By Aaditya Panigrahi, Yanjun Liu, Omri Lesser, Krishnanand Mallayya, and AI-MI Director Eun-Ah Kim (Cornell).
Source: arXiv:2606.10195