arXiv Science⌕ Search

arXiv subjects

Lizhong Fu

Publications and source records attributed to Lizhong Fu.

5 recordsLinked to original sources

Fully GPU-accelerated semistochastic Epstein-Nesbet perturbation theory

Second-order Epstein-Nesbet perturbation theory (EN-PT2) is an important tool for improving selected configuration interaction (SCI) energies, but the large external determinant space creates substantial memory and communication costs on GPU clusters. We develop a fully GPU-accelerated implementation of distributed semistochastic EN-PT2, fusing excitation generation, screening, and PT2 weighting in a single kernel traversal. Fixed-size random-phase CountSketch arrays limit memory use and replace large external determinant exchange with bounded collective summation. The implementation on one A100 completes EN-PT2 calculations 8.21-37.53 times faster than Arrow on 64 CPU cores at the same requested statistical standard error, and fixed-workload strong-scaling tests achieve 93.5% parallel efficiency on 16 GPUs across two nodes. Applications to an iron-oxo complex in a (60e, 187o) active space, iron-porphyrin references containing 200 million determinants per state, and neural-network quantum states establish the method's efficiency for perturbative corrections to large reference wave functions in the GPU era.

physics.comp-ph↗

Exact verification of Givens rotation circuits for efficient quantum phase estimation of iron-sulfur clusters

Givens rotations with angles stored in quantum registers account for much of the cost of published quantum phase estimation algorithms based on double-factorized tensor hypercontraction and spectral amplification. These rotations create superpositions and cannot be verified directly by the basis-state evaluators used to optimize classical reversible circuits. We give an exact verification method that separates out a fixed Clifford change of basis and tracks basis states and phases in the remaining circuit. Integer conditions establish correctness of the full coherent action, including measurement-based uncomputation and restoration of the phase-gradient resource state. Under our interpretation of Low et al.'s angle grid and rotation convention, simplifying the adders lowers the rotation cost relative to the published accounting; fusing the two additions into one masked addition gives an additional saving. We verify both implementations exhaustively at every precision used in the published walks and prove the fused circuit correct for all widths. For benchmarks including FeMoco and other iron-sulfur clusters, replacing only the rotation term lowers total phase-estimation Toffoli counts by 9.5-14.3% under the published error budget, with the walk operator unchanged.

quant-ph↗

Foundation Neural-Network Quantum States for Molecular Potential Energy Surfaces in Second Quantization

Second-quantized neural-network quantum states have achieved accurate molecular energies, but extending them across molecular geometries requires a shared representation of the geometry-dependent wavefunction coefficients. We introduce geometry-conditioned foundation neural-network quantum states for molecular electronic structure in second quantization. A single autoregressive model learns a family of ground states from sparse anchor geometries and provides wavefunctions at untrained geometries without further optimization. Orbital alignment matches orbital identities and transports their phases, establishing an aligned orbital basis across geometries. Frozen energies reach chemical accuracy at every untrained query geometry for N$_2$, CO, and H$_4$. On additional molecular paths, the energy-trained wavefunctions yield dipoles, quadrupoles, and natural occupations without property labels. Across three paired N$_2$ training seeds, orbital alignment lowers the mean absolute energy error over all untrained query geometries from 34-37 mHa to 0.049-0.085 mHa. At approximately 1 mHa mean absolute error, frozen evaluation reduces the per-geometry cost by $986\times$ relative to independent optimization, yielding an estimated $25.8\times$ end-to-end GPU-cost reduction on a 161-point N$_2$ grid.

physics.chem-ph↗

Scaling Neural Network Quantum States for Ab Initio Quantum Chemistry

Neural-network quantum states (NNQSs) can represent many-electron wave functions without explicitly enumerating the determinant space, but their accuracy depends jointly on model size and variational-optimization effort. Here we characterize this dependence for a physics-conditioned autoregressive NNQS trained separately on two six-molecule source benchmarks. Across eight model sizes and five optimization milestones, we find that model size and optimization steps jointly shape the energy error. The capacity advantage of larger models becomes more apparent with sufficient optimization, while the returns from additional optimization vary with model size. We capture this coupling using an interaction scaling law and quantify the cumulative compute of each evaluated configuration. The resulting error-compute Pareto frontiers provide a practical decision rule for jointly selecting model size and optimization steps under a given compute budget within the evaluated range. Furthermore, we find that this beneficial scaling trend persists during fine-tuning on held-out N$_2$. Pretrained models show decreasing error with increasing model size, with a steeper reduction following pretraining on the Hard benchmark. Together, these results place autoregressive neural quantum states within the broader landscape of empirical neural scaling and open a quantitative route toward the systematic scaling of neural quantum solvers for ab initio quantum chemistry.

physics.chem-ph↗

Clifford augmented density matrix renormalization group for \textit{ab initio} quantum chemistry

The recently proposed Clifford augmented density matrix renormalization group (CA-DMRG) method seamlessly integrates Clifford circuits with matrix product states, and takes advantage of the expression power from both. CA-DMRG has been shown to be able to achieve higher accuracy than standard DMRG on commonly used lattice models, with only moderate computational overhead compared to the latter. In this work, we propose an efficient scheme in CA-DMRG to deal with \textit{ab initio} quantum chemistry Hamiltonians, and apply it to study several molecular systems. Our numerical results show that CA-DMRG can reach higher accuracy than DMRG using the same bond dimension, pointing out a promising route to push the boundary of solving \textit{ab initio} quantum chemistry with strong static correlations.

quant-ph↗