{"id":"63545da6-de63-4151-86a5-3ff5124a60fc","arxiv_id":"2411.11476","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A partition function for lattice heteropolymers is exactly rewritten as a sign-problem-free Z2 lattice gauge theory, enabling joint Monte Carlo sampling of polymer sequences and compact structures.","lead":"This paper recasts the statistical mechanics of lattice heteropolymers as a Z2 lattice gauge theory, turning polymer configurations into gauge-field states and adding fermionic fields whose determinant removes unwanted ring polymers. The resulting Monte Carlo method samples both sequence and structure without a sign problem, and the authors demonstrate it on small two-dimensional heteropolymers and compact chains.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Ergodicity of the LGT move set is the load-bearing gap: the asserted connectivity after Eq. (5) is unproved, and zero-weight ring intermediates may separate positive-weight states.","rationale":"I have no independent objection to the formal mapping: the Gauss-law identities, the determinant computation for open chains, and the isolated-ring suppression via circulant eigenvalues are internally consistent. The load-bearing weakness is the unproved ergodicity of the MC move set, exactly as the reader identified. My stress-test sharpens it: the ring-suppression determinant does not merely remove rings from the equilibrium ensemble; it also removes them from the set of accessible intermediate configurations, because a move into a zero-weight ring state has zero acceptance probability. Thus the paper would need to prove connectivity within the positive-weight subspace, not merely within the Gauss-law subspace. The BFS enumeration I propose is small enough to be decisive and would settle whether the missing proof is a technicality or a real counterexample. The second reader concern, the weak numerical evidence for linear decorrelation scaling, is also valid but is not the central load-bearing issue: the exactness of the partition-function rewriting does not depend on it. Since the reader's conditional verdict already reflects the ergodicity gap, my assessment does not change the recommended verdict.","tokens_in":18793,"tokens_out":8925,"duration_ms":102376,"concrete_test":"On a small lattice (e.g., 4x4 or 5x5, with even M and the parameters of Eq. (A4)), enumerate all edge sets Gamma satisfying the Gauss law of Eq. (1) and hard self-avoidance (null HSA), with exactly two endpoint sites, and compute det T(Gamma). Discard all states with det T = 0. Build a directed graph whose edges are single applications of C_box (Eq. (4)) or D_ij (Eq. (5)) leading to another retained state, and require Metropolis acceptance greater than zero. Compute strongly connected components of this graph. If more than one component contains positive-weight states, the Markov chain is non-ergodic and equilibrium sampling is invalid; repeating the check at a few beta values and lattice sizes would show whether the missing ergodicity proof can be supplied or is actually false.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central algorithmic claim requires the Metropolis chain over self-avoiding single-chain states to be ergodic. The paper asserts after Eq. (5): 'Two arbitrary polymer configurations with the same number of open chains can be deformed one into the other by a combination of chain deformations and terminal displacements.' This is stated without proof or citation. Connectivity in the enlarged space of all Gauss-law states is not sufficient: the actual weight in Eq. (11) contains det T(Gamma)^M, and Appendix A shows det T = 0 for any isolated ring, via eigenvalue (A2) under the parameter choice (A4). A ring-containing configuration therefore has exactly zero Boltzmann weight, so a Metropolis step into it is never accepted. Any connecting path that must pass through a ring state is blocked. The paper does not prove that every pair of positive-weight, self-avoiding configurations can be connected through positive-weight, self-avoiding configurations using the moves C_box and D_ij. The difficulty is most acute for the headline efficiency claim, which targets >98% occupancy: in nearly full lattices, local moves are heavily constrained, and ring intermediates may be the only available routes. If disconnected positive-weight components exist, the averages in Figs. 2 and 3 are not thermal averages, and the algorithmic claim fails even though the exact rewriting is correct.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an exact mapping of the partition function of lattice heteropolymer models into a vacuum expectation value of a Z2 lattice gauge theory with both fermionic and bosonic degrees of freedom. The authors show that the Gauss-law constraint encodes polymer connectivity, that a determinant factor (det T)^M arising after integrating out M fermion flavors suppresses ring configurations when the fermion parameters are fixed as in Eq. (A4), and that a chemical potential term can tune the average chain length. They then introduce a Metropolis Monte Carlo scheme based on the local gauge-symmetry moves C_square and D_ij, and demonstrate applications to an HP-type heteropolymer and to sampling compact structures, where they report a linear scaling of the decorrelation time with chain length. The paper also discusses extensions to quantum computing.","tokens_in":19058,"tokens_out":3475,"duration_ms":39200,"significance":"If the algorithmic claims hold, this is a valuable contribution: it provides an exact, sign-problem-free rewriting of a class of lattice polymer partition functions, with a transparent mechanism for eliminating rings and a natural route toward quantum simulation. The analytical core is strong: Appendix A gives a clean proof that det T vanishes on isolated rings under the parameter choice mf = 4 gbar^2, gbar^2 = lambda^2, and Appendix B / SM Sec. D rigorously obtain det T = 4^{N-l} g^{2N} (l+1) for open chains, which is a useful, parameter-free structural result. The introduction of fermions is a construction, not a circular fit, because the ring suppression follows from the determinant structure once the parameters are fixed. The main risk is algorithmic rather than analytical: the ergodicity of the proposed Markov chain over the positive-weight subspace is asserted but not proved, and the claimed linear scaling of decorrelation time rests on limited numerical evidence. These issues are load-bearing for the Monte Carlo applications.","major_comments":[{"comment":"The statement that 'Two arbitrary polymer configurations with the same number of open chains can be deformed one into the other by a combination of chain deformations and terminal displacements' is central to the claim that the Monte Carlo algorithm samples the equilibrium ensemble, but it is asserted without proof or citation. Moreover, the relevant state space for the Metropolis chain is not the full set of Gauss-law states but the subset of self-avoiding single-chain configurations with positive weight. Because Appendix A proves that any configuration containing an isolated ring has det T = 0, such ring states have exactly zero Boltzmann weight and are never accepted by the Metropolis rule. The paper does not demonstrate that every pair of positive-weight self-avoiding configurations can be connected by a sequence of C_square and D_ij moves in which every intermediate configuration also has positive weight (i.e., is ring-free and self-avoiding). For nearly maximally compact states, where local moves are heavily constrained, the only connecting paths might pass through ring states. If the positive-weight subset is not connected, the averages shown in Figs. 2 and 3 are not thermal averages and the algorithmic claim fails. This gap needs to be fixed either by a proof of connectivity within the positive-weight subspace or by a systematic small-lattice enumeration showing that all self-avoiding single-chain states are mutually reachable through accepted moves.","section":"§From polymers to LGT, after Eq. (5)"},{"comment":"The headline claim that the MC decorrelation time 'grows only linearly with the median chain length' is inferred from a small number of simulation points, and the examples in SM Sec. F start from a specially constructed spiral Hamiltonian path. No statistical error bars are reported for the decorrelation times, and the autocorrelation function in Eq. (S4) is averaged over trajectories whose equilibration is assumed rather than demonstrated. The linear scaling is a central algorithmic result, yet on the presented evidence it could be an artifact of the chosen initial states or of insufficient statistics. I request a more systematic study: several lattice sizes, multiple random initial configurations (not just the spiral path), and either error bars on tau or a discussion of the number of independent runs and the sensitivity to the initialization.","section":"§Computational efficiency and Fig. 3"},{"comment":"The Metropolis acceptance criterion in Eq. (S1) includes the ratio of the full weights e^{L[Gamma,eta]} e^{-beta H[Gamma,eta]}, with L containing M Tr log T[Gamma]. When a proposed C_square or D_ij move produces a configuration with a ring, the acceptance probability is strictly zero since the weight vanishes. The paper does not analyze the probability that the proposed moves land on such zero-weight states, nor does it analyze the possibility that the chain gets trapped in a subset of states separated by zero-weight barriers. This is the same ergodicity concern as above, but it deserves a concrete computational check: measuring the acceptance rate of moves and testing for reducibility of the Markov chain, e.g., by comparing equilibrium distributions obtained from independent initializations.","section":"§MC Algorithm and Eq. (15)"}],"minor_comments":[{"comment":"The reference to 'Kogout' should be 'Kogut' (J. B. Kogut, Rev. Mod. Phys. 51, 659 (1979)).","section":"Introduction, Ref. [16]"},{"comment":"In the first paragraph of Appendix C, 'the NP model' should likely be 'the HP model', consistent with the main text.","section":"SM Appendix C"},{"comment":"The notation for the chain deformation operator is inconsistent: the main text and SM Fig. S6 use C_square, while SM Sec. F uses Cp. Also, the caption of Fig. 3 does not define 'median chain length' or specify how tau is estimated beyond the reference to Eq. (S4).","section":"Fig. 3 and SM Sec. F"},{"comment":"The phrase 'can we rewritten' in the sentence preceding Eq. (B5) should read 'can be rewritten'.","section":"End Matter, after Eq. (B5)"},{"comment":"The statement that 'for even values of M our LGT does not suffer from a sign problem' is unnecessarily restrictive, since the determinant for open chains in Eq. (B3) is positive for any M; please clarify whether negative signs can appear for other sectors or whether the even-M condition is merely a conservative statement.","section":"Discussion and SM Table I"}],"recommendation":"major_revision","confidential_remarks":"The ergodicity issue is the main risk to the paper's central algorithmic claims. The analytical mapping and determinant calculations are solid and should be credited. I recommend that the authors either provide a proof of connectivity within the positive-weight subspace or add a direct numerical test of ergodicity (e.g., exhaustive enumeration on small lattices) before the paper is accepted. The linear-scaling claim also needs stronger numerical support. This is a promising manuscript, but the load-bearing algorithmic claims are not yet fully supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The exact rewriting of the lattice heteropolymer partition function as a Z2 lattice gauge theory with fermions is the real contribution, and it is solid. The fermion determinant does what it claims: under mf = 4gbar^2 = 4lambda^2, det T vanishes on any isolated ring, so ring configurations have exactly zero weight. The appendix derivations (A, B, D) check out, and the determinant for an open chain depends only on its length, which makes the chemical-potential control of chain length work. The simultaneous sampling of sequence and structure through soft chemical potentials is a genuine step beyond the earlier QUBO encodings. The construction is tuned, not fitted, so there is no circularity. For even M the weight is positive, so sign-problem-free Monte Carlo is legitimate.\n\nThe soft spot is exactly the one flagged in the stress-test: ergodicity of the move set. The paper asserts after Eq. (5) that any two configurations with the same number of open chains can be connected by plaquette flips and endpoint moves, but no proof or citation is given. Because any configuration containing a ring has exactly zero weight, a Metropolis step into it is never accepted. So the connecting path must stay within the positive-weight subset. That is an assumption, and in the dense regime where the method advertises linear scaling it is not an obvious fact. If it fails, the computed averages are not thermal averages, even though the exact rewriting is correct. The numerical evidence for the linear scaling is also thin: a few points in Fig. 3, one run initialized from a specially chosen spiral Hamiltonian path, and no error bars. No code or data is released.\n\nThe core mapping is exact regardless of the MC performance. The paper deserves a serious referee. A referee should ask for a proof or a citation for the connectivity claim, or at least numerical evidence from multiple independent initializations; release code; and put error bars on the scaling measurement.\n\nWho is this for? Lattice polymer simulators and quantum-simulation people. I would send it to peer review, since the formal contribution is novel and the gap is fixable. I'd cite it for the mapping. And yes, I'd bring it to reading group—there is a good discussion here about what counts as a proven algorithmic claim.","headline":"Exact gauge-theory rewriting is solid; the unproved ergodicity of the MC moves is the real risk.","tokens_in":19642,"tokens_out":4848,"would_cite":true,"duration_ms":45830,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves an exact rewrite of heteropolymer ensembles as a sign-problem-free Z2 lattice gauge theory that Monte Carlo can sample.","keywords":["lattice heteropolymers","Z2 lattice gauge theory","Monte Carlo sampling","HP model","sign problem","quantum encoding","compact polymer conformations","ring suppression"],"falsifier":"Enumerate all self-avoiding single-chain configurations on a small two-dimensional lattice with fixed endpoints and length, and check whether the graph whose edges are accepted plaquette-flip and endpoint-displacement moves is connected; any pair of configurations with no connecting path of accepted moves would disprove ergodicity and invalidate the Monte Carlo averages. A second check is to repeat the decorrelation-time measurement starting from randomly chosen compact configurations rather than the spiral Hamiltonian path, and in three dimensions, to see whether the linear scaling survives.","tokens_in":18513,"feed_emoji":"🧬","tokens_out":9351,"duration_ms":86599,"temperature":0.7,"pith_summary":"The paper establishes an exact rewriting of the lattice heteropolymer partition function as a vacuum expectation value of a $\\mathbb{Z}_2$ lattice gauge theory, with one qubit per bond, one per site, and an auxiliary fermionic determinant that removes rings. Because the resulting Boltzmann weight is real and positive for even numbers of fermion flavors, the model has no sign problem and can be sampled by ordinary Monte Carlo in both conformational and chemical sequence space. This offers a way around the two obstacles that have made compact heteropolymer ensembles hard to simulate: growth methods that become exponentially inefficient and local moves that are overwhelmingly rejected. As demonstrations, the authors reproduce the hydrophobic collapse of a short HP-model chain and report that for compact two-dimensional configurations with more than 98% occupancy, decorrelation time grows only linearly with chain length.","feed_headline":"Polymer statistics rewritten as sign-free lattice gauge theory","feed_subtitle":"One qubit encoding samples polymer shape and chemical sequence together, with linear cost on compact chains.","key_machinery":"The load-bearing object is the fermionic determinant $\\det T(\\Gamma)$ of the tight-binding matrix $T_{ij}=\\delta_{ij}(m_f-\\bar g^2\\rho_i)-\\lambda^2\\Gamma_{ij}$. It does two jobs: it vanishes identically for any field configuration containing an isolated ring when $m_f=4\\bar g^2$ and $\\lambda^2=\\bar g^2$, so rings are expelled from the ensemble, and for the remaining open chain it depends only on chain length, $\\det T=4^{N-\\ell}\\bar g^{2N}(\\ell+1)$, so a chemical potential $\\mu$ can fix the average length. The gauge symmetry itself, the Gauss law Eq. (1), carries the argument: it encodes the polymer's continuity, and the plaquette operator $\\hat C_\\square$ and endpoint-displacement operator $\\hat D_{ij}$ generate the trial moves.","core_discovery":"The central claim is that the sum over all self-avoiding single-chain configurations, with arbitrary chemical labels, is exactly the ground-space matrix element of a $\\mathbb{Z}_2$ lattice gauge theory, Eq. (15). Polymer states are the tensor-product states satisfying the Gauss law Eq. (1); active bonds and chain endpoints are the electric and charge degrees of freedom. Integrating $M$ degenerate spinless fermion flavors coupled to the bond field yields the weight $(\\det T(\\Gamma))^M$, and the parameter choice $m_f=4\\bar g^2=\\lambda^2$ makes $\\det T$ vanish for any configuration containing an isolated ring, while for an open chain of length $\\ell$ it equals $4^{N-\\ell}\\bar g^{2N}(\\ell+1)$. The theory therefore samples exactly one open chain whose length is controlled by a chemical potential, with no sign problem for even $M$, and the same weight supports Metropolis moves that preserve the Gauss law.","pith_inferences":["Beyond the paper, the exact determinant formula $\\det T=4^{N-\\ell}\\bar g^{2N}(\\ell+1)$ suggests that on small lattices the entropy $n(\\ell)$ of chain lengths could be computed by exact enumeration, giving closed-form checks of the saddle-point length control.","Beyond the paper, the connectivity assertion could be tested as a finite-state graph problem: if some fixed-length sector is disconnected under accepted moves, equilibrium averages would be biased regardless of how fast the autocorrelation decays.","Beyond the paper, the flavor-symmetry analogy with chiral symmetry breaking invites the speculative hypothesis that protein-like sequence stability might be described by spontaneous breaking of residue 'flavor' symmetry; the authors mention this as a direction, not a result.","Beyond the paper, because the linear scaling demonstration uses a special spiral Hamiltonian path as initial condition and two-dimensional lattices, a fair stress test would be random initial compact states and three-dimensional lattices."],"forward_implications":["Simultaneous sampling of sequence and structure becomes routine: thermodynamic averages such as hydrophobic-contact counts and endpoint-distance distributions are computed directly from the same Markov chain.","Ring-free single-chain ensembles with a prescribed average length can be prepared by setting one chemical potential, without the divergent number of auxiliary constraints that earlier binary encodings needed.","For even flavor number $M$ the weight is real and positive, so standard Metropolis updates, not complex Langevin or other sign-problem remedies, are sufficient.","The observed linear growth of decorrelation time with chain length, if it generalizes, would remove the exponential slowdown that has limited sampling of maximally compact lattice polymers.","Because the encoding is qubit-based, the partition function weight could be evaluated with quantum algorithms such as block encoding or QAOA, as the paper sketches."],"supporting_citations":[{"why":"Supplies the Z2 lattice-gauge-theory formalism, including the electric operator and Gauss law, onto which the polymer continuity condition is mapped.","marker":"[16]"},{"why":"Introduced the binary/qubit encoding of lattice polymer configurations that the present work promotes to a gauge theory.","marker":"[11]"},{"why":"The earlier QUBO treatment of ring melts whose ring-suppression limitations motivate the determinant-based removal of rings.","marker":"[14]"},{"why":"Defines the two-letter HP lattice model used for the illustrative sequence and structure sampling.","marker":"[17]"},{"why":"Supplies the HP-model interaction conventions the numerical example follows.","marker":"[18]"},{"why":"Gives the circulant-matrix properties used to evaluate the loop sub-matrix eigenvalues in the ring-suppression proof.","marker":"[31]"},{"why":"Provides the exact eigenvalue formula for circulant matrices used to show the ring determinant vanishes.","marker":"[32]"},{"why":"Gives the block-determinant identity used in the supplementary proof of the open-chain determinant formula.","marker":"[33]"}],"fun_headline_variants":["Sign-free gauge theory maps heteropolymer sequence and structure","Exact LGT encoding eliminates sign problem in polymer sampling","Polymer statistics as Z2 gauge theory with linear MC cost","Qubit-based gauge theory samples heteropolymers without sign issues","Mapping to lattice gauge theory yields sign-free polymer sampling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation's equilibrium averages are valid only if the local moves can connect any two polymer configurations of the same chain length while keeping the chain self-avoiding; the paper asserts this connectivity without proof.","fun_headline_variants_meta":{"raw":{"variants":["Sign-free gauge theory maps heteropolymer sequence and structure","Exact LGT encoding eliminates sign problem in polymer sampling","Polymer statistics as Z2 gauge theory with linear MC cost","Qubit-based gauge theory samples heteropolymers without sign issues","Mapping to lattice gauge theory yields sign-free polymer sampling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000245,"raw_usage":{"total_tokens":1516,"prompt_tokens":903,"completion_tokens":613,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":519,"completion_tokens_details":{"reasoning_tokens":531}},"tokens_in":519,"tokens_out":613,"duration_ms":6885,"temperature":1.0,"reasoning_tokens":531,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T18:28:43.021240+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Enumerate all self-avoiding single-chain configurations on a small two-dimensional lattice with fixed endpoints and length, and check whether the graph whose edges are accepted plaquette-flip and endpoint-displacement moves is connected; any pair of configurations with no connecting path of accepted moves would disprove ergodicity and invalidate the Monte Carlo averages. A second check is to repeat the decorrelation-time measurement starting from randomly chosen compact configurations rather than the spiral Hamiltonian path, and in three dimensions, to see whether the linear scaling survives.","supporting_citations":[{"cited_title":"Micheletti, P","cited_arxiv_id":null,"evidence_quote":"Introduced the binary/qubit encoding of lattice polymer configurations that the present work promotes to a gauge theory."},{"cited_title":"Slongo, P","cited_arxiv_id":null,"evidence_quote":"The earlier QUBO treatment of ring melts whose ring-suppression limitations motivate the determinant-based removal of rings."},{"cited_title":"Yue and K","cited_arxiv_id":null,"evidence_quote":"Defines the two-letter HP lattice model used for the illustrative sequence and structure sampling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the HP-model interaction conventions the numerical example follows."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the circulant-matrix properties used to evaluate the loop sub-matrix eigenvalues in the ring-suppression proof."},{"cited_title":"Sinha, Eigenvalues and eigenvectors of a circulant ma- trix, in Vibration of Nearly Periodic Structures and Mis- tuned Bladed Rotors(Cambridge University Press, 2017) p","cited_arxiv_id":null,"evidence_quote":"Provides the exact eigenvalue formula for circulant matrices used to show the ring determinant vanishes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the block-determinant identity used in the supplementary proof of the open-chain determinant formula."}],"review_version":1}