{"id":"9d1909d6-cc75-4d2f-8fd0-0d2790249b6a","arxiv_id":"2501.07777","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"TeNeS-v2 extends the existing TeNeS tensor-network solver with real-time and finite-temperature simulation capabilities for two-dimensional quantum lattice systems.","lead":"TeNeS-v2 is an open-source update to the tensor-network solver TeNeS that adds real-time evolution and finite-temperature simulations for quantum spin and boson models on two-dimensional lattices. The paper documents the new algorithms and validates them on Ising and XXZ models, so researchers can run these simulations without writing their own tensor-network code.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; finite-temperature validation is narrow but the central software-capability claim is supported and the key limitations are disclosed.","rationale":"The reader's weakest assumption correctly identified the finite-temperature iTPO/simple-update approximation and its possible failure, and the paper itself flags exactly this risk. My stress-test pass found no load-bearing flaw in the central claim: the software is open source, the finite-temperature mode is benchmarked against an independent QMC method on a concrete model, and the real-time mode shows the expected entanglement-limited behavior with bond-dimension convergence. The acknowledged D=4 chirality artifact and the lack of guaranteed positivity are important caveats for users, but they are disclosed in the text and do not invalidate the claim that TeNeS-v2 enables these simulations. The only substantive reservation is that the abstract's 'various quantum spin and Bose models' is broader than the evidence provided, since all application examples are spin models and only one finite-temperature benchmark against QMC is shown. That is an overstatement of scope, but the verdict should remain ACCEPT because the central software-contribution claim is adequately supported and the limitations are transparently documented.","tokens_in":12255,"tokens_out":4271,"duration_ms":50150,"concrete_test":"Run TeNeS-v2 finite-temperature mode on the 2D S=1/2 antiferromagnetic Heisenberg square-lattice model with D=10 and compare energy, specific heat, and magnetization to QMC results for T/|J| in the range 0.2 to 2.0. Additionally, compute the eigenvalues of the resulting iTPO and check whether negative eigenvalues appear at any temperature. If severe negativity or energy below the ground-state energy occurs, the broader applicability claim would need to be qualified; if the results match QMC and positivity holds, the central claim is strengthened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that TeNeS-v2 adds working real-time and finite-temperature simulation modes to an existing iTPS/iPEPS solver. That is a software-capability claim, not a mathematical theorem, and the paper provides independent support for the finite-temperature mode through QMC comparisons in Sec. 5.2 for energy, specific heat, and magnetization of the transverse-field Ising model. The admitted restriction to simple update for iTPO and the unguaranteed positivity of the iTPO density matrix (Sec. 3.2) are genuine limitations, and the D=4 chirality singularity in Sec. 5.3 shows these limitations can produce unphysical results on a frustrated triangular-lattice model. However, the paper explicitly discloses these limitations and advises users to check bond-dimension convergence. The abstract's phrase 'various quantum spin and Bose models' is broader than the validation actually shown: no Bose model benchmark is presented and only one spin model is checked against QMC. This is an overstatement of scope, but it does not undermine the central argument that the new modes are implemented and work on the demonstrated examples.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"TeNeS-v2 is a software paper for the open-source tensor-network solver TeNeS, which targets infinite two-dimensional quantum lattice systems using iTPS/iPEPS. The new version adds real-time evolution and finite-temperature simulation modes to the previously implemented ground-state solver. The real-time mode applies Suzuki-Trotter decomposition with simple and full updates and evaluates local observables along the trajectory. The finite-temperature mode represents the thermal density matrix as a single-layer iTPO, evolves it in imaginary time from the infinite-temperature identity state using the simple update, and contracts expectation values with CTMRG. The paper also introduces multi-site observables, exemplified by scalar chirality on the triangular lattice. Validation consists of finite-temperature energy, specific heat, and magnetization for the transverse-field Ising model compared with DSQSS quantum Monte Carlo, plus a real-time quench example checked through bond-dimension convergence and energy conservation. The manuscript explicitly discloses known limitations: iTPO positive semidefiniteness is not guaranteed, full update is not yet supported for iTPO, real-time simulations are limited by entanglement growth, and the triangular-lattice example shows an unphysical singularity at D=4 attributed to the simple update.","tokens_in":12415,"tokens_out":7343,"duration_ms":75772,"significance":"If the claims hold, TeNeS-v2 is a useful community resource: it provides an accessible, open-source implementation of real-time and finite-temperature iTPS/iPEPS simulations, with reproducible example data and scripts in an ISSP repository. The finite-temperature mode is benchmarked against an independent method, DSQSS QMC, and the agreement at D=6 and D=10 for the transverse-field Ising model is genuine evidence that the implementation works for that model. The complexity analysis (single-layer contraction with O(D^6) cost when chi is proportional to D) is also helpful for users. The central software-capability claim is therefore supported for the demonstrated cases. The main weakness is the breadth of the validation: only one spin model is checked at finite temperature, no Bose-model benchmark is presented, and the real-time example relies on internal consistency checks rather than an external baseline. These caveats are, however, disclosed in the paper itself, including the explicit warning in Sec. 3.2 that iTPO may violate positive semidefiniteness. I do not see a hidden technical flaw, but the abstract and conclusion should be aligned with the narrower demonstrated scope.","major_comments":[],"minor_comments":[{"comment":"The phrase 'demonstrating TeNeS-v2's applicability to various quantum spin and Bose models' is broader than the evidence presented: the finite-temperature benchmark covers only the transverse-field Ising model and no Bose model is used in any new-feature example. Please either add an additional benchmark (for example, a hardcore-boson or Heisenberg model at finite temperature) or qualify the scope in the abstract and conclusion, for instance by saying 'spin and boson models with short-range interactions' while noting that the demonstrated examples are spin models.","section":"Abstract and Sec. 6"},{"comment":"The footnote reporting that for hx=0 a large deviation occurs when D increases above D=2 is potentially confusing and is stated without explanation. Since the main text recommends checking bond-dimension convergence, this example shows that larger D does not always improve accuracy. Please add a short explanation or a forward reference, and give concrete advice on how a user should choose D for a new model.","section":"Sec. 5.2, footnote 2"},{"comment":"The real-time demonstration uses bond-dimension dependence and energy conservation as accuracy diagnostics, but no quantitative tolerance is given. Reporting the maximum energy drift or the norm error after truncation would make the example more reproducible and would clarify what 'the discontinuity disappears' means quantitatively.","section":"Sec. 5.1, Figs. 4 and 5"},{"comment":"The statement that the step size tau in the imaginary-time evolution operator corresponds to 2tau in the actual inverse-temperature step is correct but could be made more explicit. I suggest writing the cumulative relation beta = 2 N tau for N evolution steps and stating clearly in the FT output-file description that the first column is beta, not the number of steps.","section":"Sec. 4.2, Eq. (9)"},{"comment":"There are several minor presentation issues: 'di fficult' appears in the Program Summary; the rendering 'T eNeS' is inconsistent in the abstract and body; reference [21] contains an incorrectly rendered exponent 'm = 1 2'; and in Fig. 4 the panel labels (a) and (b) are placed at inconsistent positions. These do not affect the technical content.","section":"Program Summary and references"}],"recommendation":"minor_revision","confidential_remarks":"This is a well-scoped software paper for Computer Physics Communications. The central claim is credible and the finite-temperature implementation is independently benchmarked. The main issue is that the abstract overstates the demonstrated scope, and the footnote in Sec. 5.2 about hx=0 may raise more questions than it answers. I recommend minor revision rather than acceptance in current form, and I do not see a need for additional benchmarking before publication if the authors agree to temper the scope claim. The D=4 chirality singularity and the iTPO positivity caveat are properly disclosed and should not be treated as fatal flaws."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: TeNeS-v2 is a solid software paper. The genuinely new bits are the real-time and finite-temperature modes (iTPO with simple update), plus multi-site observables, and the authors back the finite-T mode with QMC comparisons that match at D=6-10 for the transverse-field Ising model. That is real, reproducible evidence. The code is open source, samples are in a data repository, and the paper is unusually candid about failure modes: iTPO positivity is not guaranteed, and the D=4 triangular-lattice chirality calculation produces an unphysical singularity.\n\nThe real-time examples show the expected entanglement-growth limitation and D-convergence; there's no external baseline, but that is acceptable for a capability demo. The main soft spot is scope: the abstract says 'various quantum spin and Bose models,' but only one spin model is benchmarked against QMC and no Bose model is shown. That's an overstatement, not a fatal flaw. Also, the finite-T method only supports simple update, and the positivity issue means users need bond-dimension checks. The D=4 chirality artifact is concerning but disclosed, and it's exactly the kind of thing a software paper should warn about.\n\nThe paper is honest about what is an extension of existing methods and what is new. I don't see any circular reasoning; self-citations are tool references. I'd like to see a Bose-model example and maybe a second benchmark model, but those are revisions, not desk-reject material.\n\nWho is this for? Practitioners who want to run iTPS/iPEPS real-time or finite-T simulations without writing their own code. The paper deserves a serious referee; it's a genuine community resource. I'd send it to peer review.","headline":"A solid, honest software paper: the real-time and finite-temperature modes work on the demonstrated examples, and the QMC benchmarks give the finite-T claim real teeth.","tokens_in":13004,"tokens_out":1394,"would_cite":true,"duration_ms":13340,"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":"TeNeS-v2 extends a tensor-network solver to real-time and finite-temperature simulations of two-dimensional quantum lattices.","keywords":["tensor networks","iTPS/iPEPS","real-time evolution","finite-temperature simulation","quantum spin models","boson models","CTMRG","Suzuki-Trotter"],"falsifier":"Run the finite-temperature mode on a small 2D spin model whose exact thermal density matrix is known, and check whether the computed energy density ever falls below the exact ground-state energy or whether a positive observable such as $\\langle (S^z)^2\\rangle$ becomes negative. If such unphysical values persist as the CTMRG bond dimension $\\chi$ is increased without increasing the iTPO bond dimension $D$, then the single-layer iTPO representation itself, not just the contraction, is failing.","tokens_in":12033,"feed_emoji":"⚛️","tokens_out":4414,"duration_ms":43200,"temperature":0.7,"pith_summary":"The paper introduces TeNeS-v2, an open-source tensor-network program that adds real-time evolution and finite-temperature simulation to its existing ground-state solver for two-dimensional quantum lattice models. The new modes run inside the same iTPS/iPEPS framework, requiring only small input-file changes from the ground-state workflow. Finite-temperature calculations for the transverse-field Ising model are benchmarked against quantum Monte Carlo, and multi-site observables such as scalar chirality are supported. If the central claims hold, researchers can study dynamics and thermal properties of 2D spin and Bose models without writing tensor-network code themselves.","feed_headline":"TeNeS-v2 adds real-time and thermal simulation to 2D tensor-network solver","feed_subtitle":"One software package now handles ground states, quench dynamics, and finite-temperature benchmarks against quantum Monte Carlo.","key_machinery":"The central object is the iTPS/iPEPS tensor network and its mixed-state generalization, the iTPO (infinite Tensor Product Operator), whose bond dimension $D$ controls accuracy. For finite temperature, the density matrix is evolved as $\\rho(2\\tau) = e^{-\\tau H}\\rho(0)e^{-\\tau H}$ from $\\beta = 0$, with simple-update truncation, and expectation values are contracted as a single-layer square-lattice tensor network via CTMRG with environment bond dimension $\\chi$. This single-layer structure is what makes the thermal calculation computationally lighter than the double-layer contraction used for pure-state expectation values.","core_discovery":"The central claim is that TeNeS-v2 reliably performs real-time and finite-temperature simulations of infinite two-dimensional quantum lattice systems using iTPS/iPEPS. Real-time evolution is obtained by replacing the imaginary-time step $\\tau$ with $-it/\\hbar$ in the same Suzuki-Trotter update procedure used for ground states, then measuring observables at chosen time intervals. Finite temperature is obtained by imaginary-time evolving a single-layer infinite Tensor Product Operator (iTPO) starting from the infinite-temperature identity state, with CTMRG supplying the environment; because the traced iTPO is a single-layer network rather than a double layer, the thermal calculation is cheaper than a pure-state calculation at the same bond dimension. The paper demonstrates finite-temperature energy density, specific heat, and magnetization of the transverse-field Ising model against quantum Monte Carlo, and the real-time magnetization of a quenched Ising model, while noting that growing entanglement limits real-time accuracy at fixed bond dimension.","pith_inferences":["The single-layer iTPO approach could plausibly be applied to frustrated magnets or other 2D models where quantum Monte Carlo faces sign problems, but the positive-semidefiniteness caveat means users should validate results with energy bounds or independent methods.","A natural extension would be to add a purified, double-layer iTPO mode alongside the supported single-layer mode, since the paper notes the purified representation avoids unphysical densities at higher computational cost.","The observed $D=4$ scalar-chirality singularity in the triangular-lattice XXZ model suggests that the simple-update environment, rather than the iTPO representation alone, may be the limiting factor; testing against exact diagonalization on finite clusters would isolate the cause."],"forward_implications":["Users can run ground-state, real-time, and finite-temperature calculations on the same 2D lattice models by changing one mode parameter, so dynamical and thermal studies no longer require separate codes.","For the transverse-field Ising model, thermal energy density, specific heat, and magnetization computed with $D=6$ or $D=10$ match quantum Monte Carlo results, so the method is a viable alternative where QMC sign problems appear.","Multi-site observable support makes quantities like scalar chirality directly accessible, enabling tensor-network study of non-coplanar magnetic phases on frustrated lattices.","Real-time simulations at small bond dimension can show artifacts such as energy jumps, but increasing $D$ removes them, giving users a concrete convergence check for dynamics.","Real-time accuracy at fixed bond dimension is limited to short times because entanglement grows during evolution, matching known behavior of iTPS/iPEPS dynamics."],"supporting_citations":[{"why":"Supplies the real-time evolution algorithm for infinite PEPS and also discusses a purified iTPO representation for finite temperature.","marker":"[14]"},{"why":"Provides a tensor-network annealing method for two-dimensional thermal states, giving context for finite-temperature tensor-network approaches.","marker":"[20]"},{"why":"Demonstrates finite-temperature tensor-network results for the Shastry-Sutherland model, a benchmark for the new finite-temperature mode.","marker":"[21]"},{"why":"Shows an application where finite-temperature tensor-network calculations yield accurate specific heat near a quantum critical point.","marker":"[22]"},{"why":"Describes the previous version of TeNeS, whose ground-state imaginary-time evolution and input design the new version extends.","marker":"[25]"},{"why":"Gives the corner transfer matrix contraction method used for expectation values in iTPS/iPEPS and in the thermal contraction.","marker":"[26]"},{"why":"Supplies the simple-update truncation method used in ground-state and time-evolution updates.","marker":"[27]"},{"why":"Provides an early infinite-PEPS algorithm with simple-update truncation that supports the time-evolution procedure.","marker":"[28]"},{"why":"Supplies the quantum Monte Carlo reference data used for the finite-temperature Ising model comparisons.","marker":"[31]"}],"fun_headline_variants":["TeNeS-v2 adds quench dynamics and finite-T to 2D tensor solver","TeNeS-v2: real-time and finite-T for infinite 2D quantum lattices","TeNeS-v2: cheaper finite-T via single-layer iTPO"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The finite-temperature mode assumes that a single-layer iTPO, truncated by simple update and contracted by CTMRG, can represent the thermal density matrix closely enough that computed expectation values stay physical; the paper itself notes positive semidefiniteness is not guaranteed and shows an unphysical singularity in a $D=4$ triangular-lattice calculation.","fun_headline_variants_meta":{"raw":{"variants":["TeNeS-v2 adds quench dynamics and finite-T to 2D tensor solver","TeNeS-v2: real-time and finite-T for infinite 2D quantum lattices","TeNeS-v2: cheaper finite-T via single-layer iTPO"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001154,"raw_usage":{"total_tokens":4746,"prompt_tokens":873,"completion_tokens":3873,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":489,"completion_tokens_details":{"reasoning_tokens":3801}},"tokens_in":489,"tokens_out":3873,"duration_ms":29755,"temperature":1.0,"reasoning_tokens":3801,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:35:27.382518+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the finite-temperature mode on a small 2D spin model whose exact thermal density matrix is known, and check whether the computed energy density ever falls below the exact ground-state energy or whether a positive observable such as $\\langle (S^z)^2\\rangle$ becomes negative. If such unphysical values persist as the CTMRG bond dimension $\\chi$ is increased without increasing the iTPO bond dimension $D$, then the single-layer iTPO representation itself, not just the contraction, is failing.","supporting_citations":[{"cited_title":"Kshetrimayum, M","cited_arxiv_id":null,"evidence_quote":"Provides a tensor-network annealing method for two-dimensional thermal states, giving context for finite-temperature tensor-network approaches."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates finite-temperature tensor-network results for the Shastry-Sutherland model, a benchmark for the new finite-temperature mode."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the previous version of TeNeS, whose ground-state imaginary-time evolution and input design the new version extends."}],"review_version":1}