arXiv:2610.25968
Efficient Block Encodings for Ensembles of Molecular Hamiltonians
Cristian L. Cortes, Andrew J. Jenkins, William Pol, Dario Rocca, Mark Steudtner, Gian-Luca R. Anselmetti, Matthias Degroote, Nikolaj Moll, Raffaele Santagati, Michael Streif, Markus R. Hermann
Abstract
Accurately characterizing molecular properties and reactivity often requires electronic-structure calculations across an ensemble of molecular geometries. Quantum algorithms that process such ensembles coherently could address this challenge, but they require Hamiltonian block encoding circuits that combine advanced factorization techniques with sublinear scaling in ensemble size. Here, we introduce SVD-THC, which extends tensor hypercontraction to sets of Hamiltonians defined at different nuclear geometries through a singular value decomposition of the ensemble. We use SVD-THC to construct a single multiplexed block encoding circuit for all geometries in an ensemble. We then evaluate the approach on ensembles generated by varying internal rotational coordinates in 1-bromo-2-chloroethane as well as in glycine and alanine dipeptides, corresponding to established model systems for peptide conformational landscapes, with up to 5,184 geometries. Across several resolutions of the internal-coordinate grids, the Toffoli complexity of the resulting circuits grows only weakly with the number of geometries, changing by at most 12\% over grids that differ in size by up to 576$\times$. In an energy-thresholding circuit, this yields 35--200$\times$ fewer Toffoli gates than slicing together single-geometry block encodings. This approach could help make coherent processing of molecular geometry ensembles a useful primitive for fault-tolerant quantum computing.
AI Summary
Electronic-structure tasks involving potential-energy surfaces require access to Hamiltonians at many nuclear geometries. This paper asks whether their shared structure can reduce the cost of coherently block encoding an ensemble, rather than implementing independent geometry-conditioned circuits. SVD-THC constructs a joint geometry basis from one- and two-electron integral snapshots, expressing the ensemble through $K$ geometry-independent fermionic Hamiltonians. Each mode is compressed with BLISS tensor hypercontraction. A multiplexed circuit loads geometry-dependent mode weights and accommodates different THC ranks across modes. Resource estimates use this primitive in coherent energy thresholding via phase estimation and generalized quantum signal processing. Numerical tests cover 1-bromo-2-chloroethane and glycine and alanine dipeptides, with up to $5{,}184$ geometries. The relaxed ensembles retain respectively $14$, $90$, and $150$ modes. Compression is validated at every sampled geometry using CCSD correlation-energy errors below $0.6$ mHa, with exact Hartree–Fock offsets restored separately. Estimated Toffoli costs change by at most $12\%$ across grids differing by up to $576\times$. Full-grid reductions versus sliced single-geometry block encodings are approximately $35$–$201\times$; reductions versus sequential energy estimation are $10$–$60\times$. Active-volume reductions versus slicing are smaller, approximately $18$–$30\times$. The tests use minimal STO-3G bases, frozen cores, and low-dimensional torsional scans; CCSD is a proxy for quantum-resolved correlation energies. Relaxation reduces compressibility. Explicit ensemble preprocessing limits higher-dimensional applications, smoothly varying active spaces remain unresolved, and Clifford-heavy data loading limits active-volume savings. End-to-end search advantages require additional thresholding queries and depend on their constants.
Error found: The resource discussion retains a full-grid 150-mode circuit for the nine-geometry alanine subset without clearly explaining how this relates to the captions' statement that compression is redone for each grid. The observed nearly constant finite-grid costs do not establish general ensemble-size scaling without assumptions on compression ranks, normalization, and data-loading workspace. The CCSD correlation-energy tests validate a classical proxy rather than certify the compression error in exact ground-state energies returned by QPE. The QPE appendix incorrectly substitutes a product of projected oracle blocks for the projected implemented sequence and uses an unjustified state-distance-to-probability bound, leaving its claimed 5% confidence loss unproved.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.