PWO is a trust-region optimizer for autoregressive NQS that improves stability over Adam and stochastic reconfiguration methods while scaling to 1.5B-parameter models on spin systems.
Mixed citations
Title resolution pending
Mixed citation behavior. Most common role is background (50%).
21
Pith papers citing it
2,240
external citations · OpenAlex
Background
50% of classified citations
citation-role summary
background 3
method 3
citation-polarity summary
representative citing papers
Equilibrium quantum many-body methods are encoders from admissible states to represented variables, with exact decoders existing precisely when tasks are constant on encoder fibers.
First NQS variational Monte Carlo calculation of excited states in A=4 nuclei and hypernuclei, reproducing benchmarks and providing the first ab initio M1 transition strength for ^{4}_ΛH consistent with weak-coupling limit at 1.3% suppression.
Neural quantum states on K5 yield two families of approximate physical states for the Thiemann-ordered Hamiltonian constraint in Abelianized Euclidean LQG: one flat with non-zero volume (non-normalizable) and one normalizable with zero volume, close to Ashtekar-Lewandowski and Dittrich-Geiller vacua
Relativistic continuous matrix product states yield competitive variational approximations to ground state energies and observables in the phi^4, Sine-Gordon, and Sinh-Gordon models, including strongly coupled regimes.
NQS performance for TFIM ground states depends on basis choice through ground-state degeneracies and amplitude-phase uniformity, which control the convergence of multi-spin cumulant expansions.
A conditional-fidelity witness on projected small subsystems certifies entanglement, magic, and circuit complexity of generic many-body states with constant sample complexity, with fidelity certification supported by numerics.
EBMs trained with non-persistent short runs reproduce empirical data statistics via a precise dynamical process, not the equilibrium measure.
Self-attention variational wavefunctions for the 2D homogeneous electron gas up to N=169 yield energies below DMC and a converged collective-mode dispersion including a roton-like minimum.
Clifford disentanglers classified by Schmidt spectrum action reduce energy errors at fixed bond dimension in MPS simulations of molecules and improve shallow-circuit VQE calculations.
Projected Inverse Iteration reframes ground-state search for neural quantum states as an eigenvalue problem to deliver rapid, spectral-gap-insensitive convergence while retaining polynomial scaling.
Canonical mapping of quantum-dot-superconductor clusters enables neural quantum-state calculations that reveal trivial singlet, Heisenberg-like, and critical regimes with 1D gaplessness and 2D triplet states.
Transformer wave functions for the J1-J2 Heisenberg model exhibit size-independent power-law decay of V-score with compute, with the exponent decreasing as frustration increases.
A reorganized Hartree-Fock framework imposes tunable orbital locality by pairing local degrees of freedom with local solution conditions, maintaining efficient SCF optimization and competitive reaction-energy accuracy.
COO co-optimizes orbitals with TrimCI to absorb many-body correlations into the basis, cutting determinant count by orders of magnitude for iron-sulfur clusters versus localized bases or DMRG.
Neural networks represent densities in a variational extended Thomas-Fermi model, yielding binding energies within 0.5% of prior ETF results and reproducing nuclear pasta phases.
Variational autoencoders combined with symbolic regression extract physically meaningful representations and order parameters from raw quantum measurement data, revealing new phenomena such as corner-ordering in Rydberg arrays.
A discretized higher-rank gauge theory on a square lattice produces a classical Ising fracton spin liquid with preserved tensor Gauss law, but quantum perturbations induce severe Hilbert space fragmentation that blocks fractonic quantum dynamics.
NLCE+QA on trapped-ion QPU computes TFIM thermodynamic-limit energies and dispersions using ASP, VQE, and a new CX-test.
Low-energy eigenstates of a spin chain allow a neural network to reconstruct the Hamiltonian accurately, while mid-spectrum eigenstates do not, defining a spectral 'learnability' gap.
A shallow restricted Boltzmann machine variational Monte Carlo ansatz reproduces the main features of the adiabatic phase diagram and selected symmetry-broken insulating states for the one-dimensional Z2 Bose-Hubbard chain at half filling.
citing papers explorer
-
One More Time: Revisiting Neural Quantum States from a Reinforcement Learning Perspective
PWO is a trust-region optimizer for autoregressive NQS that improves stability over Adam and stochastic reconfiguration methods while scaling to 1.5B-parameter models on spin systems.
-
Full-State and Reduced-Moment Encodings: A Representation-Level View of Equilibrium Quantum Many-Body Theory
Equilibrium quantum many-body methods are encoders from admissible states to represented variables, with exact decoders existing precisely when tasks are constant on encoder fibers.
-
Neural-network excited states of $A=4$ nuclei and hypernuclei
First NQS variational Monte Carlo calculation of excited states in A=4 nuclei and hypernuclei, reproducing benchmarks and providing the first ab initio M1 transition strength for ^{4}_ΛH consistent with weak-coupling limit at 1.3% suppression.
-
Finding and characterising physical states of Euclidean Abelianized loop quantum gravity using neural quantum states
Neural quantum states on K5 yield two families of approximate physical states for the Thiemann-ordered Hamiltonian constraint in Abelianized Euclidean LQG: one flat with non-zero volume (non-normalizable) and one normalizable with zero volume, close to Ashtekar-Lewandowski and Dittrich-Geiller vacua
-
Some progress on the use of the variational method in quantum field theory
Relativistic continuous matrix product states yield competitive variational approximations to ground state energies and observables in the phi^4, Sine-Gordon, and Sinh-Gordon models, including strongly coupled regimes.
-
Basis dependence of Neural Quantum States for the Transverse Field Ising Model
NQS performance for TFIM ground states depends on basis choice through ground-state degeneracies and amplitude-phase uniformity, which control the convergence of multi-spin cumulant expansions.
-
Certifying localizable quantum properties with constant sample complexity
A conditional-fidelity witness on projected small subsystems certifies entanglement, magic, and circuit complexity of generic many-body states with constant sample complexity, with fidelity certification supported by numerics.
-
Explaining the effects of non-convergent sampling in the training of Energy-Based Models
EBMs trained with non-persistent short runs reproduce empirical data statistics via a precise dynamical process, not the equilibrium measure.
-
Accurate Self-Attention Wavefunctions at Large Scale
Self-attention variational wavefunctions for the 2D homogeneous electron gas up to N=169 yield energies below DMC and a converged collective-mode dispersion including a roton-like minimum.
-
Clifford disentanglers for entanglement reduction in molecular electronic structure simulations
Clifford disentanglers classified by Schmidt spectrum action reduce energy errors at fixed bond dimension in MPS simulations of molecules and improve shallow-circuit VQE calculations.
-
Projected Inverse Iteration: An Eigenvalue Approach to Ground-State Computation with Neural Quantum States
Projected Inverse Iteration reframes ground-state search for neural quantum states as an eigenvalue problem to deliver rapid, spectral-gap-insensitive convergence while retaining polynomial scaling.
-
Correlated States in Quantum Dot Clusters Coupled to a Common Superconductor
Canonical mapping of quantum-dot-superconductor clusters enables neural quantum-state calculations that reveal trivial singlet, Heisenberg-like, and critical regimes with 1D gaplessness and 2D triplet states.
-
Scaling Laws for Neural-Network Quantum States
Transformer wave functions for the J1-J2 Heisenberg model exhibit size-independent power-law decay of V-score with compute, with the exponent decreasing as frustration increases.
-
Approximating Hartree-Fock theory via an efficiently local reformulation
A reorganized Hartree-Fock framework imposes tunable orbital locality by pairing local degrees of freedom with local solution conditions, maintaining efficient SCF optimization and competitive reaction-energy accuracy.
-
Absorbing Many-Body Correlations into Core-Optimized Orbitals
COO co-optimizes orbitals with TrimCI to absorb many-body correlations into the basis, cutting determinant count by orders of magnitude for iron-sulfur clusters versus localized bases or DMRG.
-
Neural-Network-Based Variational Method in Nuclear Density Functional Theory: Application to the Extended Thomas-Fermi Model
Neural networks represent densities in a variational extended Thomas-Fermi model, yielding binding energies within 0.5% of prior ETF results and reproducing nuclear pasta phases.
-
Discovering quantum phenomena with Interpretable Machine Learning
Variational autoencoders combined with symbolic regression extract physically meaningful representations and order parameters from raw quantum measurement data, revealing new phenomena such as corner-ordering in Rydberg arrays.
-
Classical fracton spin liquid and Hilbert space fragmentation in a 2D spin-$1/2$ model
A discretized higher-rank gauge theory on a square lattice produces a classical Ising fracton spin liquid with preserved tensor Gauss law, but quantum perturbations induce severe Hilbert space fragmentation that blocks fractonic quantum dynamics.
-
Thermodynamic-limit dispersion relations on trapped-ion quantum hardware
NLCE+QA on trapped-ion QPU computes TFIM thermodynamic-limit energies and dispersions using ASP, VQE, and a new CX-test.
-
Information in Many-body Eigenstates: A Question of Learnability
Low-energy eigenstates of a spin chain allow a neural network to reconstruct the Hamiltonian accurately, while mid-spectrum eigenstates do not, defining a spectral 'learnability' gap.
-
Benchmarking a restricted Boltzmann machine on the $\mathbb{Z}_2$ Bose-Hubbard chain in the adiabatic hard-core regime
A shallow restricted Boltzmann machine variational Monte Carlo ansatz reproduces the main features of the adiabatic phase diagram and selected symmetry-broken insulating states for the one-dimensional Z2 Bose-Hubbard chain at half filling.