{"id":"e9e6abe9-a53e-40ca-bd1d-2fdd31c65d99","arxiv_id":"2608.10069","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A new coupled-layer construction unifies quantum product codes with coupled-layer phases and produces non-CSS stabilizer codes, including X-cube, Chamon, fermionic toric code, and Walker-Wang examples.","lead":"The authors introduce coupled-layer codes, a framework that builds quantum error-correcting codes by stacking copies of one code and coupling the layers with excitations controlled by a second code. The framework reproduces known fracton and fermionic models such as the X-cube, Chamon, fermionic toric code, and 3-fermion Walker-Wang, and it extends product-code constructions beyond CSS codes.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The disjoint-support hypothesis in Sec. VII is the genuine weak point: it is essential to the Appendix C commutativity proof, but every demonstrated example satisfies it, so the central claims about reproduced codes are not invalidated.","rationale":"The reader's weakest-assumption analysis correctly identifies the disjoint-support condition in Sec. VII as the most fragile step of the most general construction. I agree that the Appendix C proof of pairwise commutativity of the terms A(e, f~) relies on that condition: the cancellation argument in Lemma 6 needs the overlap regions for different group elements to be disjoint, which is exactly what the disjoint-support and free-permutation assumptions provide. This is a real limitation of the unified balanced coupled-layer framework, and the paper itself states that generalizing beyond this assumption is future work. However, the central claims that the paper actually demonstrates — reproducing the X-cube model, Chamon model, 3D fermionic toric code, and 3-fermion Walker–Wang model, and reducing to tensor/balanced product and e'm balanced product — all use excitations that satisfy the disjoint-support condition automatically. Single-qubit Pauli excitations have disjoint supports, and the products used in the examples are built from such single-qubit terms. Thus the concern does not invalidate the paper's positive results; it narrows the scope of the claimed unification. The reader's ACCEPT verdict remains appropriate, with moderate confidence justified by the absence of formal verification and the hand-drawn lattice identifications in the examples.","tokens_in":51251,"tokens_out":11402,"duration_ms":126026,"concrete_test":"Run a brute-force verification of the Sec. VII commutation claim on small instances with overlapping excitation supports. Enumerate all stabilizer codes H on at most three qubits, all commuting pure-type codes F on at most three qubits, and all free G-actions with onsite Cliffords as allowed in Sec. VII. For each choice of excitations e, e' with intersecting supports, compute the symplectic inner product of A(e, f~) and A(e', f~') from Eq. (11) for every pair of stabilizer representatives f~, f~'. If any pair has sign −1, the Appendix C proof fails without the disjoint-support hypothesis and the unrestricted general balanced coupled-layer construction is not a valid code switching; if all small instances commute, the assumption is sufficient but may not be necessary, substantially weakening the concern.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing premise is the disjoint-support condition imposed in Sec. VII on excitations E^{f~}, and its use in Appendix C. The code-switching terms A(e, f~) of Eq. (11) must pairwise commute for simultaneous enforcement to define a stabilizer code. Appendix C, Lemmas 3–6, proves this, and Lemma 6 is where the condition is essential: when several group elements g, h contribute to the overlap set K, the proof interleaves products over K_g and K_h and invokes cancellation of the phases φ_g and φ_h. That interleaving is justified only because the individual overlap regions (Supp U_g(e) ∩ Supp e')·g_{q2} are mutually disjoint, which follows from the supports of distinct excitations being disjoint and freely permuted by G. If overlapping excitation supports are allowed, two such overlap regions can touch in the same layer, the sign computation no longer factors, and the simultaneous condensation terms can fail to commute. The paper explicitly acknowledges this assumption in Sec. VIII, listing generalization to arbitrary excitations as future work, so it is disclosed rather than hidden. Crucially, every concrete example in the paper — X-cube, Chamon model, 3D fermionic toric code, 3-fermion Walker–Wang, and the reductions to balanced product and e'm balanced product — uses single-Pauli excitations or products whose supports are automatically disjoint and freely permuted by the group action. Therefore the concern limits only the claimed breadth of the 'unifying' balanced coupled-layer construction; it does not threaten the paper's concrete reproductions of known models.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces coupled-layer codes, a family of stabilizer-code constructions that start from a stack of copies of a first code and couple the layers with checks of a second code, condensing general Pauli excitations rather than only single-X/single-Z excitations. Three interrelated constructions are developed: (i) CSS coupled-layer codes based on mapping cones and partial gauging (Sec. III), reproducing X-cube and a 4D code; (ii) non-CSS coupled-layer codes defined through excitation algebras and algebra-preserving maps (Sec. IV), reproducing the Chamon model and the quantum XYZ product; and (iii) a generalized quotient and e'm balanced product in which a ZX-duality (Hadamard twist) accompanies the group action (Secs. V-VI), reproducing the 3D fermionic toric code and the 3-fermion Walker-Wang family. Section VII presents a balanced coupled-layer code intended to combine (ii) and (iii), with commutativity of the code-switching terms proven in Appendix C under a disjoint-support assumption. Appendix A counts logicals under stated surjectivity assumptions, and Appendix B relates e'm balancing to symplectic doubling.","tokens_in":51490,"tokens_out":14299,"duration_ms":140393,"significance":"The main contribution is a common framework in which product constructions and several topological and fracton models become instances of one condensation procedure. If the claims are correct, the framework connects code switching with gauging, mapping cones, and non-CSS topological phases. The strengths are substantial: the stabilizer groups for all the reproduced models are written out explicitly; the logical count in Appendix A is carried out under transparent assumptions; the relation to symplectic doubling is made precise; and no parameters are fitted or target results assumed. The paper is also unusually candid about its limitations, including the constant-distance 4D example, the absence of a low-dimensional example exercising both general excitations and unitary-twisted group actions, and the disjoint-support assumption in Sec. VII. The stress-test concern about that assumption is real but contained: every concrete example instantiating Sec. VII or Sec. VI uses single-Pauli or automatically disjoint excitations, so the reproduced codes are not affected; the limitation concerns the breadth of the unification claim rather than the validity of the central examples.","major_comments":[],"minor_comments":[{"comment":"In the first paragraph of Sec. II.A, 'ccohain complex' should be 'cochain complex'; please also rephrase the sentence beginning 'We always place the qubits...' so that the figure reference and the degree-zero convention are stated clearly.","section":"Sec. II.A"},{"comment":"The disjoint-support condition in Sec. VII is stated as 'the supports of e^{(\\tilde{f})} to be mutually disjoint and freely permuted by G,' which is ambiguous: Appendix C, Lemma 6 uses disjointness globally across different excitation sets E^{\\tilde{f}} and E^{\\tilde{f}'}. Please state the global version explicitly and add a remark that the plaquette excitations used in the X-cube example overlap on shared edges, so that construction is not literally an instance of the Sec. VII balanced coupled-layer construction with general excitations.","section":"Sec. VII and App. C"},{"comment":"The statement that adding the two weight-5 logicals as stabilizers results in a code with distance linear in the system size is given without proof or reference; please provide a short argument or cite a place where this is established.","section":"Sec. III.D"},{"comment":"The claim that on a three-torus the constructed code has 3 logicals if n is odd and 0 logicals if n is even is stated without derivation; please justify it by an explicit lattice calculation or a precise citation to the Walker-Wang literature.","section":"Sec. VI.C"},{"comment":"The definition of \\tilde{s}_G involves a product over operators that may become non-commuting single-qubit Pauli operators on a quotient qubit after conjugation by Hadamards; a short remark making the chosen ordering explicit and explaining why the phase ambiguity is irrelevant would improve readability.","section":"Sec. V, Eq. (10)"}],"recommendation":"minor_revision","confidential_remarks":"The paper is solid, original, and squarely in the journal's scope. The limitations that worry me most are all disclosed by the authors, and the reproductions of known models do not rest on the unproven general cases. I see no grounds for rejection; the requested changes are clarifications and small justifications rather than new technical developments."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth your time. This is the paper that actually carries out the unification that coupled-layer product-code papers have been circling: tensor and balanced product codes plus coupled-layer phase constructions land in one framework, and the e'm balanced product with an on-site Hadamard is a genuine extension, not a relabeling. The stabilizer groups are written down explicitly for X-cube, Chamon, the 3D fermionic toric code, and the 3-fermion Walker-Wang family. I checked several formulas by hand; they are consistent. Appendix A's logical count under stated assumptions is real work, and Appendix B's symplectic-doubling relation is a nice bridge for the non-CSS constructions.\n\nThe paper is honest about its limits. Section III D discloses a 4D example with constant distance, and Section VIII says no good-distance claim is made. The weakest load-bearing point is the disjoint-support condition in Section VII: the pairwise commutation of condensation terms in Appendix C, especially Lemma 6, needs it. If overlapping excitation supports are allowed, the phase-cancellation argument can fail. That is a genuine restriction on the claimed unifying scope, but the paper states it plainly and lists generalization as future work. More importantly, every concrete example in the paper satisfies the condition, so the reproductions of known models are not undermined.\n\nMinor concerns: the e'm quotient is defined up to phase and ordering; Lemma 1 shows commutativity, but I would like a sentence on whether different orderings give the same code up to Clifford equivalence rather than just phase. The fermionic-TC and Walker-Wang identifications rely on drawn lattice deformations; they are convincing but not formalized. Nothing here looks circular. The citation pattern is fine: [30] and [32] are the authors' own, but they are the natural inputs.\n\nWho this is for: people working on product-code constructions, qLDPC parameter searches, or coupled-layer topological order. A reader mainly hunting for asymptotically good codes should look elsewhere; a reader wanting a common language for these constructions will get real value.\n\nReferee call: send it out. The framework is new enough and the examples concrete enough to deserve careful refereeing, with attention to the Section VII assumption and the phase/ordering convention in the quotient.","headline":"A genuinely unifying coupled-layer framework with a new Hadamard-twisted balanced product that reproduces X-cube, Chamon, the fermionic toric code, and 3-fermion Walker-Wang; the disjoint-support assumption in the general construction is real but disclosed, and it does not threaten the concrete examples.","tokens_in":52091,"tokens_out":2127,"would_cite":true,"duration_ms":21891,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P70","81P68"],"pacs":["03.67.Pp"],"model":"deepseek-v4-flash","headline":"This paper claims that coupled-layer codes, defined by condensing arbitrary Pauli excitations across stacks via an excitation algebra and an algebra-preserving map, unify quantum product codes with coupled-layer models of topological…","keywords":["coupled-layer codes","quantum product codes","stabilizer codes","excitation condensation","mapping cone","balanced product codes","fermionic toric code","Walker-Wang model"],"falsifier":"Take the balanced coupled-layer construction with the 2D toric code as the stacked code and a repetition code as the recipe, choose a group action by translation, and pick two excitations whose supports are not mutually disjoint but are freely permuted by the group. Compute the commutator of the two code-switching terms A(e, f) and A(e', f). A nonvanishing phase would show that the disjoint-support condition is essential; universal vanishing would show Lemmas 3 through 6 are stronger than stated.","tokens_in":50984,"feed_emoji":"🧩","tokens_out":5315,"duration_ms":48571,"temperature":0.7,"pith_summary":"This paper argues that a single construction, coupled-layer codes, contains both quantum product codes and many coupled-layer models of topological phases, and extends both. The idea is to stack many copies of one code, one per qubit of a second code, then enforce new stabilizers that condense chosen Pauli excitations across layers, with the second code's checks dictating which combinations to condense. The condensation data is organized as an excitation algebra plus an algebra-preserving map, which works even when the inputs are not CSS codes. Because the formalism does not require a chain-complex description, it can produce non-CSS stabilizer codes; the authors demonstrate this by reproducing the X-cube model, the Chamon and XYZ product codes, the three-dimensional fermionic toric code, and a family containing the 3-fermion Walker-Wang model. If correct, the construction supplies a common origin for tensor products, balanced products, gauging, and mapping cones in one framework.","feed_headline":"One construction yields product codes and non-CSS phases","feed_subtitle":"Coupled-layer codes condense Pauli excitations across layers, reproducing X-cube, fermionic toric code, and Walker-Wang models.","key_machinery":"The load-bearing object is the excitation algebra: the operator algebra generated by the stabilizers of the first code together with a chosen set of Pauli excitations, paired with an algebra-preserving map that sends each generator to a Pauli operator on an auxiliary system while preserving commutation relations. This is the operator-level form of the mapping cone from homological algebra, and it replaces the chain-complex description when the input code is not CSS. The second code enters only as a recipe: for each of its stabilizers and for each qubit in that stabilizer, the corresponding image under the algebra-preserving map and the excitation in one layer are multiplied to form the code-switching term; commutativity of the second code's stabilizers guarantees these terms are compatible. In the balanced version the same data is combined with a free group action whose unitary part may be a Hadamard, which is what converts CSS inputs into non-CSS codes.","core_discovery":"The paper's central claim is that any coupled-layer condensation of Pauli excitations is specified by an excitation algebra and an algebra-preserving map, and that this data alone defines a valid stabilizer code when the second code's stabilizers commute. This formulation makes the old coupled-layer realization of tensor and balanced product codes a special case: choose single-Pauli X and Z excitations and the usual gauging maps. More importantly, the same recipe works without CSS structure, so condensing general excitations yields codes such as the X-cube and Chamon models. The paper further generalizes the quotient and balancing step by allowing the group action to include an onsite Hadamard, an e-m symmetry swapping X and Z; balancing by this symmetry turns CSS inputs into non-CSS outputs, recovering the 3D fermionic toric code and a family whose second member is the 3-fermion Walker-Wang model. The culmination is the balanced coupled-layer code, which combines general excitation condensates with permutation-plus-unitary group actions and reduces to the ordinary balanced product when the action is a free permutation and the excitations are single Pauli operators.","pith_inferences":["The disjoint-support condition on excitations looks like the true boundary of the construction: if overlapping excitation supports can still be condensed consistently, the resulting codes would go beyond the paper's framework and could have better parameters; searching for such examples is a natural next step.","The e-m quotient of the 4D toric code producing the fermionic toric code suggests a physical interpretation the authors leave implicit: compactifying the 4D toric code with an e-m duality defect inserted should realize the same fermionic phase, and a braiding or statistics calculation on the resulting model would test this.","The paper's fiber-bundle remark points toward a sheaf-theoretic reformulation in which the coupled-layer code is a fiber bundle over the second code with the first code as fiber; formalizing that view could yield a parameter-counting tool and could streamline the Appendix C calculations.","Because the authors do not analyze code parameters, an immediate testable extension is to vary the excitation algebra and algebra-preserving map to see whether the distance can be pushed beyond the constant-distance examples found in the 4D example of Section III D."],"forward_implications":["Tensor and balanced product codes are special cases of coupled-layer codes, so any property proved for the general condensation construction applies to those product families.","Condensing general excitations rather than single-Pauli operators can create fracton-type codes: the framework reproduces the X-cube model and the Chamon and quantum XYZ product codes.","Allowing a Hadamard in the balancing group action yields genuinely non-CSS codes from CSS inputs, including the 3D fermionic toric code and the family containing the 3-fermion Walker-Wang model.","The CSS coupled-layer code admits a logical-operator count under the assumptions stated in Appendix A, giving a way to compute the code dimension.","The construction supplies an explicit dictionary between gauging, partial gauging, and lattice-surgery-type operations and coupled-layer condensation, so those operations appear as special cases of one code-switching step."],"supporting_citations":[{"why":"Supplies the coupled-layer construction of tensor and balanced product codes that this paper generalizes.","marker":"[30]"},{"why":"Defines balanced product quantum codes, the target special case and baseline for the generalized e-m balanced product.","marker":"[5]"},{"why":"Defines quantum XYZ product codes, whose Chamon model example is shown to be a coupled-layer code.","marker":"[44]"},{"why":"Defines the 3D fermionic toric code, which the paper reproduces as an e-m balanced product.","marker":"[32]"},{"why":"Introduces the Walker-Wang and 3-fermion Walker-Wang models that the family of non-CSS codes is claimed to recover.","marker":"[45-47]"},{"why":"Provides the X-cube fracton model and generalized lattice gauge theory framing used as an example and comparison point.","marker":"[31]"}],"fun_headline_variants":["Coupled-layer condensation yields product and non-CSS codes","Excitation condensation unifies product and non-CSS code families","Layered Pauli condensation builds X-cube and fermionic toric code","A single recipe produces X-cube, Chamon, and Walker-Wang models","Generalized balanced product codes arise from ZX-duality"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The most load-bearing premise is that the excitations condensed in the balanced coupled-layer code have mutually disjoint supports that the group freely permutes, because Appendix C's proof that the code-switching terms commute relies on this disjointness.","fun_headline_variants_meta":{"raw":{"variants":["Coupled-layer condensation yields product and non-CSS codes","Excitation condensation unifies product and non-CSS code families","Layered Pauli condensation builds X-cube and fermionic toric code","A single recipe produces X-cube, Chamon, and Walker-Wang models","Generalized balanced product codes arise from ZX-duality"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001096,"raw_usage":{"total_tokens":4609,"prompt_tokens":1011,"completion_tokens":3598,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":627,"completion_tokens_details":{"reasoning_tokens":3507}},"tokens_in":627,"tokens_out":3598,"duration_ms":25956,"temperature":1.0,"reasoning_tokens":3507,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:14:16.355634+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the balanced coupled-layer construction with the 2D toric code as the stacked code and a repetition code as the recipe, choose a group action by translation, and pick two excitations whose supports are not mutually disjoint but are freely permuted by the group. Compute the commutator of the two code-switching terms A(e, f) and A(e', f). A nonvanishing phase would show that the disjoint-support condition is essential; universal vanishing would show Lemmas 3 through 6 are stronger than stated.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines quantum XYZ product codes, whose Chamon model example is shown to be a coupled-layer code."},{"cited_title":"Garre-Rubio, Emergent (2+1)d topological orders from itera- tive (1+1)d gauging, Nature Communications15, 7986 (2024)","cited_arxiv_id":null,"evidence_quote":"Defines the 3D fermionic toric code, which the paper reproduces as an e-m balanced product."},{"cited_title":"Gapped boundaries of (3+1)d topological orders","cited_arxiv_id":"2212.09779","evidence_quote":"Provides the X-cube fracton model and generalized lattice gauge theory framing used as an example and comparison point."}],"review_version":1}