AI and the World's Hard Problems · Chemistry & materials · arXiv · 2026-10-22
JANUS: A Multi-modal Foundation Neural Sampler for Disordered Materials
JANUS is a multimodal neural sampler that jointly explores chemical composition, atomic structure, and volume, learning directly from the system's energy without requiring pre-generated equilibrium configurations. It recovers free energy and phase behavior with about three orders of magnitude fewer energy evaluations than reference Monte Carlo sampling, reproduces experimentally observed local chemical order in multicomponent alloys, and, coupled to an evolutionary language model agent, enables efficient design of alloys with balanced optical and mechanical properties. Across semiconductors including silicon and diamond, it explores vacancies and dopants spanning 15 elements, recovers established defects such as the silicon E centre, and identifies new candidate defect pairs and triplets for quantum engineering, including S-Ti in silicon and B-[O]-O in diamond, with deep in-gap states validated by hybrid-functional density functional theory.