Pith. sign in

REVIEW 2 major objections 4 minor 11 references

Digital quantum simulation of many-body systems: Making the most of intermediate-scale, noisy quantum computers

T0 review · 2 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This thesis claims that Kibble-Zurek defect scaling in digitized quantum annealing provides an application-oriented benchmark: the number of Trotter layers that still reproduce the predicted defect-density law measures how deep a circuit a

desk verdict A solid thesis whose real contribution is the KZ-based benchmark at 133 qubits; the coherence inference is softer than the abstract implies, but the work deserves a serious referee. read the letter →

arxiv 2508.21504 v1 pith:B2M4MQLU submitted 2025-08-29 quant-ph

Alexander Miessen This is my paper · ORCID
classification quant-ph MSC 81P68 PACS 03.67.Ac03.67.Lx
keywords quantumdynamicssimulationbenchmarkingKibble-ZurekmechanismannealingerrormitigationopensystemsprobabilisticamplificationTrottercircuits
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The thesis is built around making noisy intermediate-scale quantum computers useful for simulating quantum dynamics. Its central contribution is a scalable benchmarking method: run a digitized quantum annealing circuit that crosses a quantum critical point and check how many Trotter layers still reproduce the Kibble-Zurek law, where defect density falls as the inverse square root of annealing time. Because accumulating hardware noise pushes the measured defect density off this universal scaling, the number of faithful layers becomes an intuitive quality metric that needs no classical verification and transfers to other time-evolution applications. The author demonstrates the scheme on 133 qubits, with coherent evolution up to a two-qubit gate depth of 28 and 1396 two-qubit gates. The thesis also contributes a method that reshapes characterized hardware noise into a target open-system Lindbladian via partial probabilistic error amplification, and two studies in state preparation and phase classification.

What carries the argument

The quantum Kibble-Zurek mechanism as applied to Trotterized transverse-field Ising annealing: the system is evolved under H(s) = -(1-s) Σσ^x - s Σσ^zσ^z and the density of defects (domain walls) is measured from nearest-neighbor correlators. The predicted power law ndef ∝ t_f^{-1/2} is the yardstick, and the number of Trotter layers for which the measured defect density still follows this law is the application-oriented quality metric.

What would settle it

Run the same 133-qubit benchmark on a device whose dominant errors are correlated two-qubit crosstalk not captured by the averaged depolarizing-plus-relaxation channel, and check whether the measured defect density over a range of annealing times t_f is monotonically decreasing and consistent with t_f^{-1/2} in a regime where independent simulations of the noise channel predict that coherence is already lost.

Watch

Extended reading notes

Core claim

The central claim is that in a digital quantum setting hardware noise never produces a decreasing, Kibble-Zurek-like defect density, so observing ndef proportional to t_f^{-1/2} across many Trotter steps is evidence of coherent, near-noise-free evolution. This turns quantum critical dynamics into a benchmark whose output metric is simply the number of reliably simulable circuit layers, directly transferable to applications such as quantum optimization. A corollary result, developed in Chapter 5, is that locally amplified and characterized Pauli noise can emulate a target Markovian open-system evolution with an analytically derived error bound, so that hardware noise is not only mitigated but

Load-bearing premise

The benchmark's interpretation rests on the premise that accumulating digital hardware noise can never produce a decreasing, Kibble-Zurek-like defect density, so that observing the t_f^{-1/2} scaling proves coherent evolution; this premise is supported only by a simplified averaged noise model validated at 12 qubits on one device, leaving correlated multi-qubit or crosstalk noise as a potential spoiler.

Editorial extensions

If this is right

  • The benchmark predicts, without classical verification, how deep structured time-evolution circuits can run on a given device and error-mitigation stack before noise takes over.
  • The metric is transferable: Section 4.3 shows consistency between the Kibble-Zurek benchmark and the residual energy of digitized quantum annealing applied to combinatorial optimization.
  • Because the Kibble-Zurek scaling is defined in the thermodynamic limit, the method avoids scaling issues and applies naturally to processors with more than one hundred qubits.
  • The scheme compares hardware and error-mitigation algorithms separately or in combination, giving a concrete layer count instead of an abstract fidelity.
  • Chapter 5's partial probabilistic error amplification provides a route to open quantum dynamics with a controllable, analytically derived error bound.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • My inference: the benchmark would become most valuable as a cross-generation and cross-platform standard if the same critical-dynamics experiment were repeated on devices with qualitatively different noise structures, since the current noise-model validation is carried out on a single device architecture at 12 qubits.
  • My inference: the plateau where defect density follows the Kibble-Zurek law could be used as a figure of merit for error-mitigation scaling, tracking how many additional layers each mitigation technique buys as qubit counts grow.
  • My inference: the open-dynamics emulation via amplified noise could extend to non-Markovian or spatially correlated environments if noise characterization improves beyond the sparse Pauli model, turning a hardware liability into a programmable simulation knob.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The thesis compiles four contributions around noisy intermediate-scale quantum simulation. Chapters 2–3 give a review of quantum dynamics algorithms and application outlook. Chapter 4 proposes an application-oriented benchmark based on reproducing the Kibble–Zurek defect-density scaling ndef ∝ tf^{-1/2} in digitized quantum annealing, demonstrated on IBM processors with up to 133 qubits, together with a transferability check to quantum optimization. Chapter 5 proposes partial probabilistic error amplification to engineer hardware noise into target open-system dynamics. Chapter 6 covers a hybrid electron–phonon ground-state solver and a QCNN-based phase classifier. The headline claims are coherent evolution to a two-qubit-gate depth of 28, with 1396 two-qubit gates, and a scalable benchmark that needs no classical verification.

Significance. If the benchmark is valid, it provides a genuinely scalable and intuitive quality metric for structured time-evolution circuits, which is valuable for near-term devices. The Kibble–Zurek anchor is an external textbook result, independently reproduced in statevector simulations; the 12-qubit noise-model validation and the consistency check against optimization are real, non-circular checks. The open-system PEA proposal includes an analytic error bound and numerical tests. The main risk is the load-bearing inference that observing the KZ scaling on 127–133 qubits proves coherent evolution, because the supporting noise-model study uses a simplified local depolarizing-plus-relaxation channel and does not cover correlated multi-qubit noise at the tested scale.

major comments (2)
  1. [§4.2 / §4.2.1] The central interpretive step is the claim that observing ndef ∝ tf^{-1/2} in a digital Trotterized annealing circuit is sufficient evidence of coherent evolution, because digital hardware noise 'never yields a behavior that resembles a classical thermal limit.' The manuscript itself concedes (§4.2) that a classical diffusion model can reproduce the same scaling. The rebuttal rests on noise-model simulations with Eq. (4.9), a local, averaged depolarizing-plus-relaxation channel, validated only at 12 qubits against one device (Fig. 4.2). Correlated multi-qubit noise and coherent crosstalk, which are known error sources on heavy-hex transmon processors, are not represented in this model. If such noise produced a monotonically decreasing ndef(tf) in the same tf window, the headline depth-28, 1396-gate claim would lose its interpretive foundation. Please add large-scale noisy simulations wit
  2. [§4.3 / §4.3.3] The benchmark is advertised as transferable and predictive ('the resulting quality metric is easily interpreted and transferred to other applications'). The only demonstrated transfer is the consistency check with residual energy in digitized quantum annealing for optimization, which uses the same underlying circuit family and the same device. This is a valuable first check, but it does not establish transferability to structurally different applications such as quantum chemistry or QML circuits with different gate distributions and noise sensitivity. Either add at least one structurally different second application, or narrow the abstract and introduction so that 'transferability' is presented as a hypothesis supported by one example rather than an established property.
minor comments (4)
  1. [§4.2] Please specify the fitting procedure used to extract the KZ exponent: the range of tf included in the fit, the number of circuit depths, the fitting function, and the threshold criterion for deviation from ndef ∝ tf^{-1/2}. Without this, the 'number of reliable layers' metric is not uniquely defined.
  2. [§4.2.1 / Fig. 4.2] The noise model validation would be clearer with error bars on the hardware points and the η=1 simulated curve, and a statement of the number of samples. The current statement 'in very good agreement' is not quantitatively supported in the text.
  3. [§2.2.2 / §4.1] The notations tf and ∆t are used throughout, but the relationship between Trotter time step and circuit depth could be stated once explicitly: for a fixed ∆t, the number of layers is tf/∆t. This would help the reader connect the benchmark metric to the gate depth claim.
  4. [Abstract] The phrase 'suffers from no scaling issues' is too strong; what the method achieves is no classical verification and no exponential scaling in the metric itself. Please rephrase to avoid overclaiming.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: KZ benchmark is anchored to external theory and independent statevector/noise-calibration evidence.

full rationale

The derivation chain is not circular. The KZ scaling (Eq. 4.5) is an externally established theoretical result, and the paper independently verifies the expected finite-size scaling regimes in noiseless statevector simulations (Fig. 4.1c). The hardware benchmark consists of measuring ndef(tf) on real devices and comparing the observed slope to that external prediction; the threshold depth is an experimental finding, not a parameter fitted to the claim. The noise-model simulations used to argue that digital gate noise does not produce a classical-thermal-like decreasing ndef (Section 4.2.1, Eq. 4.9) are based on calibration data (T1, T2, gate errors) from ibm_sherbrooke; the η=1 model is checked against a 12-qubit hardware curve, so the model is anchored to independent calibration and device data rather than to the 127–133-qubit KZ result. The paper itself flags the classical diffusion counterexample and addresses it by simulation, which is a robustness concern at large scale, not a definitional equivalence. The transferability to optimization uses the same digitized-QA circuit family, so it is a consistency check of the metric rather than an independent prediction; this is a scope limitation but not a circular reduction. Thesis self-citations are republication of peer-reviewed work and do not constitute load-bearing unverified support.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claims rest on standard quantum physics (Schrodinger and Lindblad equations), on the Kibble-Zurek prediction for a 1D transverse-field Ising chain, and on three distinctly load-bearing modeling choices: (1) that averaged per-gate depolarizing-plus-relaxation noise faithfully represents what large IBM devices do, and that this noise class can never reproduce a decreasing KZ-like defect density (validated at 12 qubits only); (2) that twirled hardware noise can be truncated to a sparse second-order Pauli model and then reshaped into a target Lindbladian for open-system simulation; (3) that the VQE, NGS, and QCNN parameters fitted during optimization genuinely solve the stated physics or classification tasks. The free parameters are calibration inputs or chosen discretizations and fit forms; no new physical entities are postulated.

free parameters (6)
  • Device noise model parameters (eta, T1, T2, e1q, e2q, ero) = T1=266.37 us, T2=178.71 us, e1q=1.25e-3, e2q=1.10e-2, ero=2.41e-2; eta in {0.01,...,10}
    Eq. (4.9): averaged calibration data from ibm_sherbrooke, used to argue digital noise cannot mimic KZ scaling. The argument holds only within this noise class.
  • Trotter time step dt = dt=0.5 for hardware benchmarks; 0.01/0.1 in statevector scans
    Chosen discretization (Section 4.1, Fig. 4.2). The benchmarked 'simulable depth' is measured in layers of this fixed dt, so results depend on the choice.
  • Annealing schedules A(s), B(s) = linear: A(s)=1-s, B(s)=s
    Section 4.1, Eq. (4.3): chosen for simplicity; defect density scaling depends on schedule derivatives.
  • PEA extrapolation fit function = single or double exponential in noise gain G
    Sections 2.3.5, 5.1.2: exact G-dependence is a linear combination of exponentials; the truncated fit form introduces model error acknowledged in the text.
  • Sparse Pauli noise model truncation order = second order (up to two-qubit Pauli channels)
    Section 5.1.1: higher-order correlated channels are dropped; this sets what the Chapter 5 method can learn and reshape.
  • VQE and QCNN parameters = optimized by classical routines during training
    Section 6.1 (energy minimization) and 6.2 (label minimization): the phase diagrams and classification accuracies are outcomes of fitted parameters.
assumptions (6)
  • standard math Time evolution is governed by the Schrodinger equation (closed) and the Markovian Lindblad master equation (open)
    Eqs. (2.1), (2.5), (2.6); underlies all simulated dynamics in Chapters 4 and 5.
  • domain assumption Kibble-Zurek mechanism: defect density in a 1D TFIM crossing its critical point scales as ndef ~ tf^{-1/2}
    Eq. (4.5); the benchmark's target prediction. Reproduced in noiseless statevector simulations (Fig. 4.1c) for N=10-20, which supports but does not prove the scaling at 127-133 qubits.
  • domain assumption Hardware noise in digital circuits cannot mimic the decreasing KZ scaling; a persistent t^{-1/2} law implies coherent evolution
    Sections 4.2-4.2.1, Fig. 4.2. The authors flag that a classical diffusion model can mimic KZ scaling, and rebut with 12-qubit noisy simulations using a simplified averaged noise model (Eq. 4.9). This is the key interpretive premise for the headline claim.
  • domain assumption After Pauli twirling, gate noise is a Pauli channel that can be truncated at second order without losing the effects that matter
    Sections 2.3.3 and 5.1.1; Ref. 175. Four-qubit and higher correlated noise is neglected; load-bearing for the PEA-based method in Chapter 5.
  • domain assumption Reshaped hardware noise can faithfully emulate a chosen system-environment interaction (target Lindbladian)
    Section 5.2, including locally amplified noise (5.2.1). The method's validity depends on sufficient overlap between characterized noise sectors and the target jump operators; tested numerically in 5.2.3-5.2.4.
  • domain assumption The Hubbard-Holstein model with the NGS-VQE ansatz and the QCNN training data capture the physics or phases under study
    Sections 6.1.1-6.1.2 and 6.2.2-6.2.3. Results are only as good as the model choice and the variational or trained ansatz classes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Digital quantum simulation of many-body systems: Making the most of intermediate-scale, noisy quantum computers." pith.science (2026). https://pith.science/paper/B2M4MQLU

@misc{pith2026250821504,
  author       = {Pith},
  title        = {Pith review of: Digital quantum simulation of many-body systems: Making the most of intermediate-scale, noisy quantum computers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B2M4MQLU}},
  note         = {Machine review of arXiv:2508.21504}
}
read the original abstract

Quantum mechanical problems are among the hardest to simulate and, in some cases, remain intractable even for the most powerful computers. Quantum computing has emerged as a new technological platform to address such challenges, with rapid advances in recent years. Yet, current quantum devices remain noisy and limited in scale. Hence, it is essential to identify classically hard problems of practical interest and tractable with existing quantum devices. Among potential applications, the real-time simulation of quantum systems is one of the most promising to deliver an early, practical quantum advantage. This doctoral thesis is therefore centered around simulating quantum dynamics on quantum devices. We first present an overview of the most relevant quantum algorithms for quantum dynamics, highlighting respective advantages and limitations. Further, we identify relevant problems within quantum dynamics that could benefit from quantum simulation in the near future. Second, we propose a method for benchmarking hardware and error mitigation algorithms that is based on well-understood theoretical results and suffers from no scaling issues. The resulting quality metric is intuitive and transferable to other applications. We successfully implement the scheme on up to 133 qubits, demonstrating coherent evolution up to a two-qubit gate depth of 28, featuring 1396 two-qubit gates, before noise becomes prevalent. Third, we propose a novel variant of probabilistic error amplification to implement open quantum dynamics, relying on characterizing and altering hardware noise to mimic the system-environment interaction under study. Lastly, we present two studies on state preparation and phase classification: first, a hybrid algorithm to prepare ground states of electron-phonon systems; second, a quantum machine learning-based approach to distinguish phases in previously prepared quantum states.

Figures

Figures reproduced from arXiv: 2508.21504 by the authors.

Figure 2.1
Figure 2.1. Development timeline of quantum algorithms for quantum dynamics. Non￾exhaustive timeline visualizing the development of decomposition and variational quantum algorithms for quantum dynamics and respective further developments. Each point represents the date of first appearance, typically as arXiv preprint. All corresponding references are summarized in [PITH_FULL_IMAGE:figures/full_fig_p038_2_1.png] view at source ↗
Figure 2.2
Figure 2.2. Dynamical decoupling and Pauli twirling. a Dynamical decoupling sequences can be entered during idle times in circuits, typically in ladders of two-qubit gates as they have the longest execution times. Here, an XY4 sequence with spacing τ is entered to keep the top qubit from idling and decohering. b Pauli twirling randomly samples Paulis (indicated by different colors) around noisy gate layers, typically two-qubit … view at source ↗
Figure 2.3
Figure 2.3. Zero noise extrapolation and probabilistic error amplification. a Schematic representation of zero noise extrapolation. Note that this does not represent any real data and only serves to illustrate the concept. An expectation value is measured at several noise gains Gi . Those values are fitted and the exact expectation value at zero noise is approximated via extrapolation of the fit to G = 0. b In probabilistic err… view at source ↗
Figures from the paper (19 more)
Figure 3.1
Figure 3.1. Figure 3.1: Rating of time evolution algorithms and applications. Assessment of the practica￾bility of quantum algorithms for quantum dynamics and progress of applications grouped into sub-areas. For each metric (row), we provide a color-coded rating, where dark green, green, ye…
Figure 4.1
Figure 4.1. Figure 4.1: Application-oriented benchmarking of quantum simulations. a Benchmarking experiments give detailed insight into device characteristics such as error rates and decoher￾ence times. As devices scale, however, it is desirable to have a benchmarking method that resembles …
Figure 4.2
Figure 4.2. Figure 4.2: Density of defects scaling in the presence of simulated hardware noise. a All curves show simulated Trotterized time evolution of a 12-qubit chain with ∆t = 0.5. The noise-free reference curve (pink line) represents a statevector simulation. The hardware curve (purpl…
Figure 4.3
Figure 4.3. Figure 4.3: Density of defects scaling in 1D from 100-qubit circuits. Hardware results com￾paring the density of defects scaling on subsets of 100 qubits on ibm_torino (left) and ibm_sherbrooke (right) employing different levels of EMS. Each point corresponds to one additional T…
Figure 4.4
Figure 4.4. Figure 4.4: Density of defects scaling on a heavy-hexagonal lattice from 133- and 127-qubit circuits. a Statevector results comparing the density of defects scaling in a periodic chain, a heavy-hexagonal lattice consisting of two heavy-hex cells (2×1), and a 7×3 square lattice, …
Figure 4.5
Figure 4.5. Figure 4.5: Hardware results of the residual energy dependence on time step and spectral gap. a Spectra relative to the ground state of three different instances of disorder in a 12-qubit periodic chain with coupling coefficients uniformly sampled from Ji j ∈ [−1,1]. The instanc…
Figure 4.6
Figure 4.6. Figure 4.6: Hardware results of the residual energy dependence on the time step. a Residual energy of a heavy-hexagonal 133-qubit graph with coupling coefficients uniformly sampled from Ji j ∈ [−1,1], averaged over 105 samples obtained from QA on ibm_torino with fixed time steps…
Figure 5.1
Figure 5.1. Figure 5.1: Scaling of the PEA extrapolation error. a Extrapolated expectation values 〈Z Z〉 as a function of the total number of circuit samples M distributed over several noise gains. The expectation value is measured after five time steps as described in the main text. Every p…
Figure 5.2
Figure 5.2. Figure 5.2: Noise model extrapolation Clifford vs. non-Clifford. a Time evolution of a dissipative Ising model without transverse field. These results correspond to the ones shown in [PITH_FULL_IMAGE:figures/full_fig_p093_5_2.png]
Figure 6.1
Figure 6.1. Figure 6.1: Hybrid quantum-classical electron-phonon solver. a The hybrid quantum algorithm iterates between a variational quantum eigensolver (VQE) for the electronic part and a non-Gaussian solver (NGS) for the phonon part of the many-body ground state. The quantum circuit str…
Figure 6.2
Figure 6.2. Figure 6.2: Phase diagram of the one-dimensional Hubbard-Holstein model. a, b Charge N(π) and spin S(π) structure factors for the ground state of a 6-site Hubbard-Holstein model, simulated with the NGS-VQE algorithm as a function of u for fixed a λ = 1.5 and b λ = 3.5. The charg…
Figure 6.3
Figure 6.3. Figure 6.3: Scaling behavior of the hybrid electron-phonon solver. a Simulation error for the ground-state energy ENGS−VQE of the circuit ansatz in [PITH_FULL_IMAGE:figures/full_fig_p103_6_3.png]
Figure 6.4
Figure 6.4. Figure 6.4: Transpilation of variational circuit to linear connectivity. a Required connectivity for the HVA circuit of a 1D Hubbard lattice in b (same as in [PITH_FULL_IMAGE:figures/full_fig_p105_6_4.png]
Figure 6.5
Figure 6.5. Figure 6.5: Influence of noise on the performance of the hybrid electron-phonon solver. a Phase diagram of the 1D Hubbard-Holstein model. The highlighted points were used for the noise simulations of panel b. b Relative error of the converged ground state energy for three distin…
Figure 6.6
Figure 6.6. Figure 6.6: Architecture of the quantum convolutional neural network. Structure of a prototypical QCNN circuit, consisting of convolutional layers (CL), pooling layers (PL), and a fully connected layer (FCL). Within the QDL framework, the input state ρin is a quantum state, prep…
Figure 6.7
Figure 6.7. Figure 6.7: Schwinger model phase diagrams. Average electric field in the ϑ-m/g plane, where ϑ acts as an external gauge field. From left to right, the plots show system sizes and lattice spacings of a Ns = 4,a = 1, b Ns = 6,a = 1, c Ns = 8,a = 2. The red line highlights the par…
Figure 6.8
Figure 6.8. Figure 6.8: Variationally prepared ground states of the Schwinger model. Results of varia￾tionally preparing the ground states of the Schwinger model within the parameter ranges m/g corresponding to the highlighted regions in [PITH_FULL_IMAGE:figures/full_fig_p115_6_8.png]
Figure 6.9
Figure 6.9. Figure 6.9: QCNN outputs for ground state phase recognition in the Schwinger model. Labels output by the QCNN after training on ground states of the Schwinger model with Ns = 8, ag = 2, ϑ = π, averaged over 20 trainings. The open blue and filled orange markers represent training…
Figure 6.10
Figure 6.10. Figure 6.10: QCNN outputs for classifying confinement in a Z2 gauge theory. Labels output by the QCNN after training on time-evolved states of the Z2 gauge theory with Ns = 2 at T = 2, averaged over 20 trainings. The open markers represent training data, while filled markers rep…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

11 extracted references · 5 canonical work pages

  1. [474]

    Self-verifying variational quantum simulation of lattice models

    C. Kokail et al. “Self-verifying variational quantum simulation of lattice models”. In: Nature 569.7756 (2019), pp. 355–360. DOI : 10.1038/s41586-019-1177-4

    doi:10.1038/s41586-019-1177-4 2019

  2. [475]

    Über das Paulische Äquivalenzverbot

    P . Jordan and E. Wigner. “Über das Paulische Äquivalenzverbot”. In: Zeitschrift für Physik 47.9 (1928), pp. 631–651. DOI : 10.1007/BF01331938

    doi:10.1007/bf01331938 1928

  3. [476]

    Finite-representation approximation of lattice gauge theories at the continuum limit with tensor networks

    B. Buyens et al. “Finite-representation approximation of lattice gauge theories at the continuum limit with tensor networks”. In: Phys. Rev. D 95 (9 2017), p. 094509. DOI : 10.1103/PhysRevD.95.094509

    work page doi:10.1103/physrevd.95.094509 2017

  4. [477]

    SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python

    P . Virtanen et al. “SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python”. In: Nature Methods 17 (2020), pp. 261–272. DOI : 10.1038/s41592-019-0686-2

    doi:10.1038/s41592-019-0686-2 2020

  5. [478]

    Floquet approach to $\mathbb{Z}_{2}$ lattice gauge theories with ultracold atoms in optical lattices

    C. Schweizer et al. “Floquet approach to Z2 lattice gauge theories with ultracold atoms in optical lattices”. In: Nature Physics 15.11 (2019), pp. 1168–1173. DOI : 10.1038/s41567- 019-0649-7. arXiv: 1901.07103 [cond-mat.quant-gas]

    work page Pith review arXiv 2019

  6. [479]

    Confinement in a Z2 lattice gauge theory on a quantum com- puter

    J. Mildenberger et al. “Confinement in a Z2 lattice gauge theory on a quantum com- puter”. In: Nature Physics 21.2 (2025), pp. 312–317. DOI : 10.1038/s41567-024-02723-6

    doi:10.1038/s41567-024-02723-6 2025

  7. [480]

    Provably efficient machine learning for quantum many-body prob- lems

    H.-Y. Huang et al. “Provably efficient machine learning for quantum many-body prob- lems”. In: Science 377.6613 (2022). DOI : 10.1126/science.abk3333

    doi:10.1126/science.abk3333 2022

  8. [481]

    Improved machine learning algorithm for predicting ground state properties

    L. Lewis et al. “Improved machine learning algorithm for predicting ground state properties”. In: Nature Communications 15.1 (2024). DOI : 10.1038/s41467-024-45014- 7

    work page doi:10.1038/s41467-024-45014- 2024

Show all 11 references
  1. [482]

    Quantum Convolutional Neural Networks are (Effectively) Classically Simulable

    P . Bermejo et al. “Quantum Convolutional Neural Networks are (Effectively) Classically Simulable”. In: arXiv preprint (2024). DOI : 10.48550/arXiv.2408.12739. arXiv: 2408. 12739

    2024 doi arXiv

  2. [483]

    Entropy of entanglement and correlations induced by a quench: Dy- namics of a quantum phase transition in the quantum Ising model

    L. Cincio et al. “Entropy of entanglement and correlations induced by a quench: Dy- namics of a quantum phase transition in the quantum Ising model”. In: Phys. Rev. A 75 (5 2007), p. 052321. DOI : 10.1103/PhysRevA.75.052321

    2007 doi

  3. [484]

    Dynamics and correlations at a quantum phase transition beyond Kibble-Zurek

    K. Roychowdhury, R. Moessner, and A. Das. “Dynamics and correlations at a quantum phase transition beyond Kibble-Zurek”. In: Phys. Rev. B 104 (1 2021), p. 014406. DOI : 10.1103/PhysRevB.104.014406. 150

    2021 doi

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.