Transformer Atomic Cluster Expansion: TRACE

Original · EN

Designing machine-learning interatomic potentials involves achieving the precise representation of complex many-body interactions alongside the efficiency required for scalable molecular dynamics. We introduce Transformer Atomic Cluster Expansion (TRACE), an energy-conserving architecture that combines atomic cluster expansion density correlations with local multihead cross-attention. The correlations form an O(3)-equivariant state for each center, which queries tensorial neighbor features that remain fixed functions of species and geometry. No learned state is passed between atoms. On a laptop MacBook-M1, we train and test TRACE for polymorphic cesium lead iodide, liquid water, and intramolecular methyl migration against experiments. For cesium lead iodide, TRACE reproduces the r²SCAN+rVV10 ordering of four polymorphs and gives a classical edge-sharing hexagonal non-perovskite(δ) to corner-sharing cubic perovskite(α) Gibbs-free-energy crossing ≃580K near the experimental observations of ≃600K. By employing enhanced sampling to cross high energy barriers, the same TRACE potential successfully captures the δ-to-α perovskite transformation without any reinforcement learning. A water potential trained on a reduced set of CCSD(T) configurations places the first oxygen--oxygen maximum at 2.85 Å, compared to the experimental value of 2.80 Å. For the gas-phase methyl migration in 2,2-dimethylisoindene, umbrella sampling yields an activation free energy of 27.92±0.03 kcal mol⁻¹, in close agreement with the experimental measurement of 29.2±1.1 kcal mol⁻¹. Across these diverse benchmarks, a single unified architecture successfully captures multi-species crystallization, liquid structures, phase diagrams, and chemical reactivity.