{"id":"80cad9ab-8014-4d85-ae88-57cb4b0d972a","arxiv_id":"2510.14781","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A documented C++/Python package for finite-temperature quantum Monte Carlo of the toric code in parallel fields, adding high-temperature/zero-field updates and reproducing the known phase boundary.","lead":"ParaToric is a new open-source software package that uses continuous-time quantum Monte Carlo to simulate the toric code under parallel X and Z magnetic fields at finite temperature. Its new Monte Carlo updates target an ergodicity problem at high temperature and zero off-diagonal field, and its benchmarks reproduce a known topological phase transition near h_c≈0.33.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The new all-time-axis flip updates (§4.1) are claimed to restore ergodicity at high T and zero off-diagonal field, but no detailed-balance proof, acceptance-ratio derivation, or ergodicity argument is given, and the benchmark at h≈0.33 does not exercise this regime.","rationale":"The reader's weakest_assumption identifies exactly the same point: the new all-time-axis flip updates are asserted to restore ergodicity but come with no derivation or benchmark in the target regime. This is load-bearing because the abstract's claim of improved applicability rests on these updates, and the phase-boundary benchmark at h≈0.33, λ=0.2, T=1/L does not test them. The exact-diagonalization check proposed here directly probes whether the updates are correct and ergodic in the previously inaccessible regime. Since the reader already flagged this as the core weakness and issued a CONDITIONAL verdict, my stress-test does not move the verdict; it confirms that the condition should be met before the extension claim is accepted.","tokens_in":23079,"tokens_out":10087,"duration_ms":85329,"concrete_test":"For a small periodic square lattice (e.g., 3×3 or 4×4 unit cells), run ParaToric at λ=0 (zero off-diagonal field) and high temperature (e.g., T=1/β with β=1–2), and compare the QMC estimates of energy, <σ^x_l>, and anyon density against exact diagonalization of Eq. (1). If the QMC results agree within error bars, the ergodicity/detailed-balance concern is resolved; if not, the new updates are incorrect or insufficient.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central novelty is the claim that two new updates—flipping the spin on the entire imaginary-time axis on one bond, or on a plaquette/star (§4.1)—restore ergodicity at high temperature and at zero off-diagonal field. However, §4.1 gives only a one-paragraph description: no proposal probabilities, no derivation of the acceptance ratio, no detailed-balance verification, and no argument that the updates connect all relevant sectors. The benchmark in §4.4.4 is at low temperature (T=1/L) and nonzero λ=0.2, so it does not exercise the regime the new updates target. The paper itself advises (Sec. 4.3.9) that added features should be benchmarked against analytical results or other numerical methods; no such benchmark is provided for the new updates. If the global-flip updates violate detailed balance or fail to connect the relevant spin sectors, then the abstract's claim of enabled ergodicity and improved applicability is unsupported, while the phase-transition reproduction could still pass. This is the most load-bearing weak point because it concerns the paper's stated extension and the correctness of potentially all results in the newly claimed regime.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents ParaToric, a C++/Python software package for continuous-time quantum Monte Carlo (CT-QMC) simulations of the toric code in a parallel field. It implements the Wu–Deng–Prokof'ev algorithm on square, triangular, honeycomb, and cubic lattices with open or periodic boundaries, and it introduces two new updates that flip a spin on the entire imaginary-time axis on one bond or on a plaquette/star. The paper documents the API, installation, observables, diagnostics, and benchmarks, including a reproduction of the known topological transition at h_c(λ=0.2)≈0.33. The central claimed novelty is that the two global-flip updates restore ergodicity at high temperatures and zero off-diagonal field, but this claim is asserted without a detailed-balance proof or a numerical test in the target regime.","tokens_in":23438,"tokens_out":6115,"duration_ms":55867,"significance":"If the algorithmic extension is correct, ParaToric would be a useful, interoperable finite-temperature solver for the extended toric code, with attractive features: multi-language bindings, snapshot export to GraphML, autocorrelation diagnostics, and bootstrap error analysis. The phase-transition benchmark is a genuinely nontrivial validation: no parameter is fitted, and the reported h_c is consistent with independent literature [5]. Credit is also due for the explicit thermalization and autocorrelation guidance, which is often missing in codebase papers. However, the only advertised algorithmic novelty — the two full-axis flip updates — is not supported by any derivation or by a benchmark in the regime they are designed for. Since the abstract and conclusion rest on this extension, the paper currently cannot be accepted as a validated codebase publication.","major_comments":[{"comment":"The two new full-axis flip updates are described only verbally. There is no statement of the proposal distribution, no Metropolis acceptance-ratio derivation, and no proof that the combined seven-update set satisfies detailed balance or connects all spin sectors. The statement that 'the spin at imaginary time 0=β cannot be flipped' is an ergodicity assertion, not a proof. This matters because the abstract and conclusion claim that these updates 'enable ergodicity at large temperatures and at zero off-diagonal field.' The paper's own §4.3.9 advises benchmarking new features, but no benchmark in the target regime is provided. Please add a derivation (including the change in the total integrated diagonal energy and the proposal probabilities) and a numerical validation against exact diagonalization or a sector-connectivity check in the high-temperature/zero-field regime.","section":"§4.1, Monte Carlo Updates"},{"comment":"This benchmark does not exercise the regime the new updates target. It uses T=1/L (low temperature) and λ=0.2 (finite off-diagonal field), where the original five Wu–Deng–Prokof'ev updates are expected to be ergodic. A clean reproduction of h_c(λ=0.2)≈0.33 therefore does not support the claim that the global-axis flips restore ergodicity at high temperature or zero off-diagonal field. Add benchmarks at small β and at h=0 or λ=0, with acceptance ratios and autocorrelation times, and compare against exact diagonalization or high-temperature series for small systems.","section":"§4.4.4, Topological phase transition"},{"comment":"The manuscript text does not provide a repository URL, version tag, or DOI for the ParaToric source code. For a codebase submission, reproducibility of the benchmarks and inspection of the update implementation are central. Please add a clear Code Availability statement and, if possible, provide reviewer instructions for reproducing the benchmark figures with a fixed seed.","section":"Code availability"}],"minor_comments":[{"comment":"The text says 'the only parameter that the user can change is the number of bootstrap resamples N_between_steps.' The parameter name should be N_resamples, not N_between_steps.","section":"§4.2.3, Error bars"},{"comment":"The sentence 'The optimal choice for N_samples is the integrated autocorrelation time' appears to be a typo; presumably the optimal thinning interval N_between_samples should be compared with τ_int.","section":"§4.3.6, Choosing N_between_samples"},{"comment":"The abstract mentions 'smooth open boundaries' while the body only lists 'open' boundaries. Please align the terminology.","section":"§2.2 / Abstract"},{"comment":"The phrase 'spin at imaginary time 0=β' is confusing; it should read 'τ=0, which is identified with τ=β by periodicity.'","section":"§4.1"},{"comment":"The runtime benchmark uses a single run per system size and notes non-identical laptop conditions. This is acceptable as an indicative benchmark, but it would be clearer to label the table as a single-run timing estimate rather than a robust performance measurement.","section":"§4.4.3, Run-time"}],"recommendation":"major_revision","confidential_remarks":"The main issue is the unsubstantiated algorithmic novelty. I would not recommend acceptance until the detailed-balance and ergodicity question for the global-flip updates is resolved, either by a proof or by a targeted benchmark in the high-temperature/zero-field regime. The h_c benchmark itself is not circular because the reference value appears in independent ref. [5], but the authors should avoid relying on ref. [9] for the 'known' value. Also ensure that the source code and benchmark scripts are accessible to reviewers, since the manuscript as submitted does not include a repository link."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a legitimate software paper. The package implements the Wu-Deng-Prokof'ev CT-QMC algorithm and adds multiple lattices, open boundaries, C/C++/Python interfaces, HDF5/GraphML snapshot output, and hysteresis workflows, then validates itself against the known h_c≈0.33 transition on the square lattice. If you work on toric-code QMC, this is a genuinely usable tool and it deserves a serious referee.\n\nThe new algorithmic content is in Sec. 4.1: two updates that flip the spin on the entire imaginary-time axis on one bond or on a plaquette/star, claimed to restore ergodicity at high temperature and zero off-diagonal field. The stress-test concern is fair. The text gives no proposal probabilities, no acceptance-ratio derivation, no detailed-balance check, and no argument that these updates connect the relevant sectors. The paper itself tells users to benchmark new features against analytical results or exact diagonalization (Sec. 4.3.9), but no such benchmark is provided for these updates. The benchmark in Sec. 4.4.4 runs at T=1/L with λ=0.2, which does not exercise the regime the updates are meant to unlock. So the claim of enabled ergodicity at high T and zero field is currently unsupported.\n\nI don't want to overweight this. The core benchmark—reproducing the known phase boundary—holds up. The code reports autocorrelation times, thermalization diagnostics, and an energy consistency check; nothing points to a bug in the standard updates. The single-shot benchmarks are a minor weakness; a code paper should ideally show a couple of seeds and error bars on timings, but it's not disqualifying. The circularity worry doesn't land: the critical value is taken from the literature and no parameter is fitted to it.\n\nThe missing analysis is addressable. Referees should ask for the detailed-balance derivation of the two new updates, or at least a test in the high-T/zero-field regime against exact diagonalization or a different QMC implementation. If the updates are wrong, the abstract overclaims, but the rest of the package remains useful for the regimes where the original algorithm is already ergodic.\n\nRecommendation: accept for peer review. The code is substantial, documented, and MIT-licensed; the referees should require the missing justification before publication.","headline":"Well-built, well-documented CT-QMC code for the toric code; the new global-flip updates are the load-bearing novelty and they lack the detailed-balance/ergodicity documentation that the journal should require.","tokens_in":23859,"tokens_out":2534,"would_cite":true,"duration_ms":21473,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"ParaToric is a continuous-time quantum Monte Carlo package for the toric code in parallel X and Z fields, extended with two new updates that keep sampling ergodic at high temperature and zero off-diagonal field.","keywords":["toric code","continuous-time quantum Monte Carlo","parallel field","topological order","quantum spin liquid","phase boundary","percolation order parameter","finite-temperature simulation"],"falsifier":"On a small (e.g., 4×4) square lattice at h=0, λ=0, and high temperature, run a chain from a staggered initial configuration and check whether the spin at τ=0 ever flips. Then compare the acceptance ratio reported for the full-axis update with the exact Metropolis ratio obtained by enumerating the diagonal-energy change for each possible configuration; a mismatch, or a chain that remains stuck in one τ=0 sector, would falsify the ergodicity claim.","tokens_in":23010,"feed_emoji":"⚛️","tokens_out":9190,"duration_ms":70849,"temperature":0.7,"pith_summary":"The paper introduces ParaToric, a finite-temperature quantum Monte Carlo package for simulating the toric code in parallel X and Z fields. Its central claim is that the established continuous-time quantum Monte Carlo algorithm can be extended with two new full-imaginary-time-axis spin-flip updates, and that these updates restore ergodicity in regimes—high temperature and zero off-diagonal field—where the original updates cannot flip the spin at the periodic imaginary-time boundary (0=β). If this is right, the package opens previously inaccessible parts of the phase diagram to sign-problem-free simulation and provides a reusable tool for studying topological order, confinement, and snapshot generation. The paper supports the claim by reporting that percolation, Fredenhagen-Marcu, and staggered-imaginary-time order parameters all reproduce the known square-lattice phase boundary at h_c(λ=0.2)≈0.33.","feed_headline":"New updates keep toric-code Monte Carlo ergodic at zero field","feed_subtitle":"It reproduces the known phase boundary and extends sampling into high-temperature, zero-field regimes.","key_machinery":"The central object is the continuous-time quantum Monte Carlo representation of the extended toric-code partition function, in which worldlines in imaginary time are generated by off-diagonal spin flips. On top of the original five update types, ParaToric adds two full-imaginary-time-axis updates: one flips a single bond's spin for the entire imaginary-time interval, the other flips all spins on a plaquette (in the X-basis) or a star (in the Z-basis) for the entire interval. These updates move the spin at the boundary τ=0=β, which is otherwise frozen in high-temperature and zero-off-diagonal-field regimes, thereby restoring connectivity of the sampled configuration space. Their stated cost i","core_discovery":"ParaToric implements the original continuous-time quantum Monte Carlo algorithm for the toric code in parallel X and Z fields and extends it with two updates that flip an entire bond's spin, or an entire plaquette's (or star's) spin, along the full imaginary-time axis. These updates are introduced to cure a specific ergodicity failure: at high temperature and at zero off-diagonal field, the spin at the periodic boundary (0=β) cannot be flipped, so the original update set cannot move between certain sectors. The paper asserts that the new updates make the Markov chain ergodic in those regimes, cost little because they only change the cached total diagonal energy, and also shorten autocorrelat","pith_inferences":["Because the full-axis updates only involve diagonal energy, the same construction should transfer to other sign-problem-free stabilizer-plus-field models—e.g., generalized Ising gauge theories with longer-range diagonal couplings—provided the new update set is checked for balance.","The percolation and Fredenhagen-Marcu observables need only equal-time snapshots, so the snapshot output could be used directly to train neural-network state classifiers or to benchmark tensor-network results.","The zero-field regime is exactly where the topological phase is most stable, so if the new updates are truly ergodic there, ParaToric could probe ground-state physics at large L without the usual restriction to finite off-diagonal fields.","A natural next step would be to measure the off-diagonal Fredenhagen-Marcu string operators and Rényi entropies, which the paper lists as requiring major code changes; the current package's clean interface makes such extensions a concrete test of its design."],"forward_implications":["Users can simulate the perturbed toric code on square, triangular, honeycomb, and cubic lattices with periodic or open boundaries at finite temperature.","The new updates extend the reach of the algorithm to high temperatures and to h=0 or λ=0, regimes that were previously problematic or inaccessible for ergodic sampling.","Snapshot output in both X and Z bases lets the package generate training and benchmarking data for lattice-gauge-theory simulators, cold-atom quantum simulators, and error-correction studies.","The reported near-independence of runtime from β means very low temperatures can be reached with only logarithmic search overhead.","Reproducing the known h_c≈0.33 boundary validates the implementation on a global phase-transition observable, not just local quantities."],"fun_headline_variants":["ParaToric's bond updates fix zero-field ergodicity in toric code","Toric-code Monte Carlo: new updates restore ergodicity at zero field","Ergodic toric-code sampling at zero field with ParaToric","Full-imaginary-time flips beat zero-field barrier in toric code","ParaToric makes zero-field toric-code sampling fully ergodic"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the two full-imaginary-time-axis flip updates satisfy detailed balance and connect all configurations that the original updates cannot reach; the paper asserts this but does not derive the acceptance probabilities or prove ergodicity.","fun_headline_variants_meta":{"raw":{"variants":["ParaToric's bond updates fix zero-field ergodicity in toric code","Toric-code Monte Carlo: new updates restore ergodicity at zero field","Ergodic toric-code sampling at zero field with ParaToric","Full-imaginary-time flips beat zero-field barrier in toric code","ParaToric makes zero-field toric-code sampling fully ergodic"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000377,"raw_usage":{"total_tokens":1817,"prompt_tokens":690,"completion_tokens":1127,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":434,"completion_tokens_details":{"reasoning_tokens":1028}},"tokens_in":434,"tokens_out":1127,"duration_ms":9034,"temperature":1.0,"reasoning_tokens":1028,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T09:30:13.597888+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a small (e.g., 4×4) square lattice at h=0, λ=0, and high temperature, run a chain from a staggered initial configuration and check whether the spin at τ=0 ever flips. Then compare the acceptance ratio reported for the full-axis update with the exact Metropolis ratio obtained by enumerating the diagonal-energy change for each possible configuration; a mismatch, or a chain that remains stuck in one τ=0 sector, would falsify the ergodicity claim.","supporting_citations":[],"review_version":2}