Electronic structure theory is an invaluable tool in chemistry due to its predictive power as well as its ability to explain complicated chemical phenomena. The usefulness of electronic structure methods is limited only by the computational complexity, which grows much more rapidly than the size of the system. We will demonstrate that the quantities used to describe the interaction of electrons as well as the quantum mechanical wavefunction contain much less information than their size would indicate. We will show how to exploit this fact to build compressed representations of the data. Using these reduced representations, the efficiency of many electronic structure methods can be greatly improved. Specically, in the case of the highly accurate coupled cluster singles and doubles method (CCSD) the computational complexity can be reduced by two orders from O(N6) to O(N4).

Reduced scaling in electronic structure theory via tensor hypercontraction

248th National Meeting (2014)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.