arXiv:2609.25834
Nullspace-guided Adaptive Bootstrap of Quantum Many-body Systems
Xu-Cheng Wang, Yang Qi
Abstract
We introduce a nullspace-guided adaptive (NGA) bootstrap method that improves the energy lower bounds of quantum many-body ground states by refining the bootstrap basis in a dynamic and incremental way. At each iteration, the optimized moment matrix reveals a nullspace of saturated positivity directions, which is intuitively interpreted as annihilators of the approximate ground-state subspace. The NGA bootstrap then prunes operators with small nullspace leverage and grows the basis along descendants of these null directions. By applying the NGA bootstrap to the transverse-field Ising chain, we obtain nearly exact energy lower bounds because the algorithm automatically discovers the eigenoperator structure in terms of Jordan-Wigner fermions from a minimal local bootstrap basis. For the Hubbard chain, it improves upon state-of-the-art energy lower bounds by up to two orders of magnitude, reaching errors ranging from $10^{-3}$ down to $10^{-5}$ in the strongly correlated regimes. We further show that the NGA framework can be used to improve the certified two-sided bounds on general observables. In addition, the bootstrap error decreases approximately as a power law with increasing computational resources. These results suggest that our method provides a practical and scalable route toward accurate bootstrap of general quantum many-body systems.
AI Summary
Many-body bootstrap methods give ground-state energy lower bounds through semidefinite constraints on truncated moment matrices, but their accuracy and cost depend strongly on the operator basis. This paper introduces nullspace-guided adaptive (NGA) basis selection: after each SDP solve, it removes operators with low participation in the optimized moment matrix’s nullspace and adds missing operator components generated by Hamiltonian commutators of null directions. The motivation connects saturated positivity constraints to approximate ground-state annihilators and dual sum-of-squares certificates; basis selection remains heuristic. Tests cover finite periodic Ising and Hubbard chains. Starting from the local basis $\{I,X,Z\}$, NGA reaches numerical precision for the critical transverse-field Ising chain at $L=32$, discovering strings corresponding to Jordan–Wigner fermion bilinears. A nonintegrable longitudinal-field example also shows decreasing errors. For the Hubbard chain at $U/t=4$, the half-filled $L=10$ energy-density lower-bound gap falls from a previously reported $5\times10^{-3}$ to $6.5\times10^{-5}$. Reported gaps at $L=16,32$, including doping, range from $5.7\times10^{-4}$ to $4.9\times10^{-3}$, benchmarked against exact diagonalization or DMRG. Energy-constrained observable optimization bounds the $L=10$ double occupancy between $0.103900$ and $0.104239$. Demonstrations are restricted to finite one-dimensional systems. Adaptive trajectories depend sensitively on initialization, pruning makes bases nonnested, and observed power-law error reduction is empirical. SDP runtime and memory are identified as primary bottlenecks: larger Hubbard runs reach 66–75 hours per step on 16 cores, with up to 1.5 TB allocated. Extensions to higher dimensions, finite temperature, and the thermodynamic limit are proposed but not demonstrated.
Error found: The assumed PSD realization used to interpret Hubbard moment-nullspace operators as ground-state annihilators is impossible for a non-tight bound, because the imposed zero particle-number variance would place that realization entirely in the target sector, where its energy cannot fall below the exact ground-state energy.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 13 pages, 5 figures