Search papers, labs, and topics across Lattice.
The paper introduces stochastic tensor contraction, a method to reduce the computational cost of high-order tensor contractions common in ab initio quantum chemistry. This approach is applied to coupled cluster theory with perturbative triples, achieving near mean-field scaling for chemical accuracy. Benchmarks demonstrate an order-of-magnitude improvement in computation time and error compared to local correlation approximations, with reduced sensitivity to system properties.
Stochastic tensor contraction slashes the cost of gold-standard quantum chemistry calculations, rivaling mean-field theory in speed while surpassing it in accuracy.
Many computational methods in ab initio quantum chemistry are formulated in terms of high-order tensor contractions, whose cost determines the size of system that can be studied. We introduce stochastic tensor contraction to perform such operations with greatly reduced cost, and present its application to the gold-standard quantum chemistry method, coupled cluster theory with up to perturbative triples. For total energy errors more stringent than chemical accuracy, we reduce the computational scaling to that of mean-field theory, while starting to approach the mean-field absolute cost, thereby challenging the existing cost-to-accuracy landscape. Benchmarks against state-of-the-art local correlation approximations further show that we achieve an order-of-magnitude improvement in both total computation time and error, with significantly reduced sensitivity to system dimensionality and electron delocalization. We conclude that stochastic tensor contraction is a powerful computational primitive to accelerate a wide range of quantum chemistry.
