{"id":"5a97a41f-549c-4bb8-8e10-a53bfeee1262","arxiv_id":"2505.06212","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A ZX-calculus framework unifies linear, ternary-tree, and local fermion-to-qubit encodings and yields a direct algorithm for the binary matrix of any ternary tree mapping.","lead":"This paper draws fermion-to-qubit mappings as diagrams in the ZX-calculus, a graphical language for quantum circuits, and shows that three popular families of such mappings appear as the same kind of diagram. It offers a new way to see when two mappings are equivalent and to compute the matrix that defines a mapping directly from a ternary tree.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1's proof that the ZX encoder matches the tree mapping relies on an unstated uniqueness theorem from [11]; if its hypotheses exclude some tree shapes, Algorithm 1's claimed output may not be the tree's encoding matrix.","rationale":"The reader's weakest assumption points to the same external uniqueness theorem of Chiew et al. [11], and my reading agrees that this is the most load-bearing unsupported step in the central ternary-tree claim. The paper's Theorem 2 and Algorithm 1 ultimately rely on Theorem 1, whose proof uses the external theorem to move from 'same set of Pauli strings' to 'same unitary map'. Because the theorem is cited only as an arXiv preprint and its hypotheses are not stated, the paper is not self-contained on this point; if the uniqueness result has hidden restrictions, the equivalence and the algorithm's correctness could fail. I do not see an internal inconsistency: spot-checking Algorithm 1 on small trees gave matrices whose encoded Majorana strings match the ternary-tree Pauli strings, so the construction appears sound. The other flagged issue, Lemma 4 and its deferred proof, affects the Hamiltonian-term diagrams of Section 3.2 but is not load-bearing for the ternary-tree-to-linear-encoding claim. The correct disposition remains CONDITIONAL: the authors should either restate and verify the external theorem's hypotheses or supply a self-contained proof, and they should specify how Algorithm 1 handles absent branches. These are addressable gaps, not demonstrated errors.","tokens_in":24646,"tokens_out":57952,"duration_ms":555802,"concrete_test":"Enumerate all ternary trees on up to n=5 labelled nodes, compute the encoding matrix by Algorithm 1, and compare it with the matrix obtained by the constructive procedure in Chiew et al. [11] (or by directly solving the vacuum-preserving Pauli-string pairing). Additionally, extract the precise hypotheses of the uniqueness theorem from [11] and verify that every enumerated tree satisfies them, including degenerate branches with no labelled descendants. If the matrices coincide and all trees satisfy the hypotheses, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is in Theorem 1's proof: it asserts that verifying the correct set of Pauli strings suffices because 'there is a unique product-preserving mapping for the set of ternary tree Pauli strings, as proved by Chiew et al. [11]'. The hypotheses and exact statement of that uniqueness theorem are not restated, and the cited theorem itself already proves the paper's central equivalence. This matters because the paper quotes uniqueness only 'up to symmetries such as fermionic braids and Pauli relabelling'; if those symmetries are nontrivial, the same Pauli-string set can correspond to different Fock-basis encoders. The ZX encoder fixes one assignment by pushing Jordan-Wigner Majoranas through the diagram, and Algorithm 1 is claimed to produce the corresponding matrix, but the proof does not show this assignment matches the specific product-preserving pairing of Miller et al. A failure of the external theorem's hypotheses for some tree shapes or labelings would therefore break the claimed correctness of Algorithm 1 and the graphical equivalence. My own spot checks of Algorithm 1 on small trees (root with one X child; root with X and Y leaves; the Jordan-Wigner comb) are consistent with the claimed matrices, so this is a proof-completeness gap rather than a demonstrated counterexample.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a ZX-calculus framework for fermion-to-qubit mappings. It establishes a correspondence between linear Fock-basis encodings and unitary phase-free ZX-diagrams (Section 3), gives a translation from ternary tree mappings to scalable ZX-diagrams (Section 4), and claims to graphically prove that every ternary tree mapping is a linear encoding. The main new algorithmic contribution is Algorithm 1, which constructs the binary encoding matrix ET directly from a ternary tree without first enumerating Pauli strings. The paper also derives controlled ZXW diagrams for electronic Hamiltonian terms under arbitrary linear encodings (Section 3.2) and presents graphical encoder/stabilizer descriptions of the E-type and square-lattice auxiliary-qubit local encodings (Section 5). The overall aim is to unify the operator-centric, Fock-state, and stabilizer perspectives on fermion-to-qubit mappings in a single graphical language.","tokens_in":24775,"tokens_out":6849,"duration_ms":72550,"significance":"If the main theorems are fully established, this paper delivers a useful synthesis: Algorithm 1 is a concrete, directly implementable procedure that lets practitioners read off encoding matrices for ternary-tree mappings without computing Pauli strings, and the ZX representation connects operator-centric tree mappings to CSS/stabilizer descriptions and to CNOT circuits. The paper independently reproduces the recent equivalence result of Chiew et al. using a genuinely different, diagrammatic proof route, which is a valuable cross-check, and the worked examples for Jordan-Wigner, parity, and Bravyi-Kitaev encodings are clear sanity checks. The local-encoding diagrams, if verified, would offer a compact unified view of encoder isometry, stabilizers, and interaction geometry. However, the current proof-completeness gaps described below mean that the central equivalence and the local-encoding tiling claim are not yet fully certified. I regard these gaps as fixable within the scope of a revision rather than as fundamental errors.","major_comments":[{"comment":"The proof of Theorem 1 verifies only that the set of Pauli strings obtained by pushing Jordan-Wigner Majorana operators through the encoder matches the ternary tree's Pauli strings, and then invokes Chiew et al. [11] for uniqueness of the product-preserving mapping 'up to symmetries such as fermionic braids and Pauli relabelling.' The hypotheses and exact statement of that uniqueness theorem are not restated, and 'up to symmetries' may not be enough to identify the specific linear encoding: fermionic braids and Pauli relabellings act nontrivially on the Fock-basis matrix, so the same Pauli-string set can correspond to different encoding matrices. The proof therefore does not establish that the ZX encoder (and hence Algorithm 1) realizes the specific product-preserving pairing of Miller et al.; a failure of the external theorem's hypotheses for some tree shapes or labelings would invalidate the claimed correctness of Algorithm 1 and Theorem 2. I ask the authors to state the external theorem precisely and to prove (or cite a proof) that the diagrammatic assignment coincides with the pairing used by Chiew et al., not merely that the Pauli-string sets agree.","section":"Section 4.3, Theorem 1 (main text and Appendix A.4)"},{"comment":"The proof of Lemma 4 is deferred to an unpublished Master's thesis [1] and a manuscript in preparation [2]. Lemma 4 is load-bearing for the controlled-diagram composition used in Propositions 6-10 of Section 3.2, so the claimed controlled diagrams for electronic Hamiltonian terms are not established within this paper. Either include a complete proof (which appears to be a short diagrammatic argument) or explicitly reformulate the affected propositions so that they do not depend on an unpublished result.","section":"Appendix A.2, Lemma 4"},{"comment":"The paper asserts that the plaquette encoder 'tiling this as in Figure 6 gives a ZX-diagram for the square lattice AQM on lattices of any size' and that the encoder reproduces the hopping terms of Steudtner and Wehner. No proof is given for the tiling, for the boundary stabilizers, or for the claim in Remark 1 that any choice of linearly independent logical operators is valid and equivalent up to a unitary on the logical qubits. Since this section is presented as a contribution rather than a conjecture, these assertions need a verification argument, or at minimum a precise reference, before they can be accepted as part of the paper's results.","section":"Section 5.2, Eq. (35) and Figure 6"}],"minor_comments":[{"comment":"The conjugation steps in Eq. (21) are difficult to parse because the two sides of the equality do not clearly display the direction of conjugation; please rewrite the display with an explicit 'E O E†' form.","section":"Section 3, Eq. (21)"},{"comment":"Algorithm 1 assumes the fixed node-ordering convention described in Section 4.2. The remark that arbitrary labelings can be handled by a permutation is not reflected in Theorem 1 or in Algorithm 1; please state explicitly how the permutation is absorbed into the encoding matrix ET.","section":"Section 4.2-4.3, Algorithm 1"},{"comment":"The proofs use abbreviations such as S1, S2, RCopy, PT, GCopy, matmult, inv, Z,X, fuse, and OCM without defining them or pointing to the corresponding rules in Figure 1 and the scalable ZX literature. A short rule-name table would make the derivations reproducible.","section":"Appendix A.1, Propositions 2 and 4"},{"comment":"In Eq. (33), the notation 'M inc_1' and 'M inc_2' is introduced as incidence matrices of graphs given by biadjacency matrices M1 and M2, but the subsequent diagram label uses 'M1 M2'; please make the notation consistent.","section":"Section 5, Eq. (33)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript leans on the independent theorem of Chiew et al. for a load-bearing step; asking the authors to restate that theorem and prove the assignment match is essential. I would also request that Lemma 4's proof be included or that the affected Hamiltonian-term claims be reworked, since dependence on unpublished manuscripts is not acceptable for core lemmas in a journal submission. The new Algorithm 1 and the unifying diagrammatic perspective are the main novel contributions beyond the already-established equivalence, and I believe a focused revision can bring the paper to publishable quality."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Best understood as a graphical repackaging of a known theorem, not a new equivalence result. What earns its keep is Algorithm 1 and the ZX translation itself. The central claim—ternary trees yield linear encodings—was already proved by Chiew et al., and the authors say so. The novelty is in presentation: ternary tree mappings become phase-free ZX diagrams that keep the tree's connectivity, and once you have that, the encoding matrix can be read off recursively. I spot-checked Algorithm 1 on small cases (root with one X child, root with X and Y leaves, the Jordan-Wigner comb) and it gives the claimed matrices. The diagrams for Jordan-Wigner, parity, and Bravyi-Kitaev also reduce to the expected matrix arrows. The local encoding section is more exploratory, but reading stabilizers off the diagram is a nice trick and it does link the stabilizer and isometry pictures.\n\nThe soft spots are real but mostly presentation-level. Theorem 1's proof says checking the Pauli strings is enough because of Chiew et al.'s uniqueness result 'up to symmetries such as fermionic braids and Pauli relabelling'. Those symmetries are not stated precisely, and the proof does not show the ZX encoder picks the same representative as the product-preserving pairing. This is a proof-completeness gap, not a counterexample; my spot checks are consistent. Lemma 4, used for the Hamiltonian-term diagrams, is deferred to unpublished manuscripts [1,2]; for a journal version that needs to be proved or replaced with a published reference. And the claim that the square lattice auxiliary-qubit-mapping plaquette tiles to any lattice size is asserted without proof. These are the points a referee should push.\n\nThe citation pattern is honest: the central equivalence is credited to Chiew et al., and self-citations are only for background ZXW rules. No overclaiming.\n\nThis paper is for people working on fermion-to-qubit mappings who want a common language to compare encodings, and for ZX practitioners looking for a new application domain. It deserves a serious referee; with the gaps above addressed it would be a solid contribution. I would engage with it.","headline":"A useful graphical repackaging of a known ternary-tree result; the new algorithm and diagrammatic translations are the real contribution, but the proof has some unstated dependencies.","tokens_in":25426,"tokens_out":2300,"would_cite":true,"duration_ms":23493,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P68","68Q12"],"pacs":["03.67.-a","03.67.Lx"],"model":"deepseek-v4-flash","headline":"Using the ZX-calculus, the paper proves that every ternary-tree fermion-to-qubit mapping is a linear encoding and unifies linear, tree-based, and local encodings in a single graphical language.","keywords":["fermion-to-qubit mapping","ZX-calculus","scalable ZX-calculus","ternary tree mapping","linear encoding","local encoding","stabilizers","Bravyi-Kitaev transform"],"falsifier":"Take a concrete ternary tree, compute its encoding matrix with Algorithm 1, and independently compute the matrix from the Pauli strings produced by the cited pairing scheme; any mismatch between the two matrices would falsify Theorem 2.","tokens_in":24342,"feed_emoji":"🌳","tokens_out":7581,"duration_ms":70512,"temperature":0.7,"pith_summary":"This paper argues that the ZX-calculus, a graphical language for quantum maps, offers a unified way to see the many different fermion-to-qubit mappings as the same kind of object. It establishes that a linear encoding of the Fock basis—one that sends occupation-number basis states to qubit basis states through an invertible binary matrix—is exactly a unitary phase-free ZX-diagram. Its main theorem is that the encoder of any ternary-tree mapping is such a phase-free diagram, so every ternary tree yields a linear encoding, and the binary matrix can be read directly from the tree by a recursive algorithm. The same framework represents local encodings with auxiliary qubits as Clifford ZX-diagrams whose connectivity mirrors the fermionic Hamiltonian's interaction graph and whose stabilizers are visible at a glance. A sympathetic reader would care because this turns a scattered zoo of constructions (Jordan-Wigner, parity, Bravyi-Kitaev, ternary trees, local codes) into one picture where comparisons and calculations can be done graphically.","feed_headline":"Ternary tree mappings are linear encodings, in pictures","feed_subtitle":"A graphical proof reads a tree's binary encoding matrix straight off the branches, unifying linear and local encodings.","key_machinery":"The central object is the phase-free fragment of the scalable ZX-calculus: a network of Z- and X-spiders with zero phases, extended with bold register wires and matrix arrows that stand for bipartite graphs of spiders. A matrix arrow labelled by a binary matrix $A \\in \\mathbb{F}_2^{m\\times n}$ implements the linear map $|x\\rangle \\mapsto |Ax\\rangle$, which is exactly what a linear encoding does to Fock basis states. The paper's load-bearing move is a local replacement rule: a ternary tree node with $a$, $b$, and $c$ descendants along its X, Y, and Z branches becomes a phase-free diagram containing the anti-diagonal matrix $F$, spliced into the wires for those branches. Rewriting the whole tree diagram to its phase-free normal form collapses it to one matrix arrow, and that rewriting is packaged as Algorithm 1, which recursively assembles the encoding matrix from the subtree matrices $E_X$, $E_Y$, and $E_Z$. Correctness is checked by pushing Jordan-Wigner Majorana strings through the diagram, which recovers the tree's Pauli strings; the local-encoding half of the paper instead uses isometries in the Clifford/stabilizer fragment, whose graph-state normal form reveals stabilizers.","core_discovery":"The paper's central claim is that the three standard presentations of fermion-to-qubit mappings—binary-matrix linear encodings, ternary trees with a Majorana-pairing scheme, and local encodings built from stabilizers—are all captured by one fragment of the ZX-calculus. Phase-free ZX-diagrams that are unitary correspond precisely to linear encodings of the Fock basis: a diagram's normal form is a matrix arrow labelled by the encoding's binary matrix, and any such diagram can be rewritten as a CNOT circuit. Ternary tree mappings translate node-by-node into phase-free ZX-diagrams, with each node replaced by a small diagram containing an anti-diagonal matrix arrow; pushing the Jordan-Wigner Majorana operators through the encoder reproduces exactly the Pauli strings the tree generates. Therefore every ternary tree mapping is a linear encoding, and a recursive reading of the tree (Algorithm 1) outputs its encoding matrix without first constructing Pauli strings. For local encodings, the encoder is an isometry in the stabilizer fragment, represented with graph-state normal forms, and the same diagrams display both the stabilizer group and the interaction geometry of the Hamiltonian.","pith_inferences":["If the correspondence is as tight as claimed, deciding whether two fermion-to-qubit mappings are equivalent could be reduced to rewriting one ZX-diagram into the other, giving a decision procedure that avoids comparing long lists of Pauli strings.","Algorithm 1 suggests an inexpensive search over ternary trees: enumerate tree shapes, compute encoding matrices directly, and rank mappings by operator weight or connectivity without ever materialising the Pauli strings.","Because local encodings now live in the same stabilizer-fragment language as quantum error-correcting codes, code-design tools such as graphical normal forms for stabilizer codes could be repurposed to construct fermion-to-qubit mappings with desired error-correction properties.","The future-work direction of bosonic systems, if carried out in infinite-dimensional ZX-calculus, would test whether the same phase-free normal-form reasoning extends to boson-to-qubit encodings."],"forward_implications":["Every ternary-tree fermion-to-qubit mapping can be implemented as a CNOT circuit, because the phase-free ZX-diagram produced from the tree reduces to CNOT circuits.","The encoding matrix of any ternary-tree mapping can be computed directly from the tree's shape by Algorithm 1, without first deriving the Pauli strings.","All one- and two-body terms of an electronic Hamiltonian obtain controlled ZXW diagrams under any linear encoding, so encoded Hamiltonians can be derived and simplified graphically.","Local encodings can be presented as stabilizer-fragment isometries whose diagrams carry the interaction graph, the stabilizers, and the encoder in a single picture.","The framework subsumes Jordan-Wigner, parity, and Bravyi-Kitaev transforms as special cases of one graphical normal form."],"supporting_citations":[{"why":"Supplies the external theorem that ternary tree transformations are equivalent to linear encodings and that the product-preserving Majorana pairing is unique, which Theorem 1 invokes to verify the translation.","marker":"[11]"},{"why":"Defines ternary tree mappings and the product-preserving Majorana-pairing scheme whose output the ZX translation is required to reproduce.","marker":"[37]"},{"why":"Introduces the scalable ZX-calculus with bold register wires, divide and gather nodes, and matrix arrows, the notation used for every diagram in the paper.","marker":"[9]"},{"why":"Establishes the phase-free normal form for interacting bialgebras, which is what reduces a phase-free diagram to a single matrix arrow.","marker":"[5]"},{"why":"Gives the pushing-through-isometry technique that lets Jordan-Wigner operators be commuted past encoder diagrams to obtain encoded Pauli operators.","marker":"[25]"},{"why":"Presents the E-type and square-lattice auxiliary-qubit local encodings whose encoder diagrams and stabilizers are analysed in Section 5.","marker":"[50]"},{"why":"Introduces the ZXW-calculus and W node used to build controlled diagrams of electronic Hamiltonian terms.","marker":"[47]"}],"fun_headline_variants":["ZX-calculus shows all fermion encodings are one","Ternary trees are linear encodings, graphically","One ZX framework for fermion-to-qubit mappings","Graphical unification of fermion-to-qubit encodings","Fermion encodings: binary, tree, local—all ZX"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is an external theorem, cited as [11], that every ternary tree admits a unique product-preserving way of pairing its Majorana operators; if that uniqueness fails for some tree shape or labelling, the ZX diagram claimed to be the tree's encoder could instead describe a different mapping.","fun_headline_variants_meta":{"raw":{"variants":["ZX-calculus shows all fermion encodings are one","Ternary trees are linear encodings, graphically","One ZX framework for fermion-to-qubit mappings","Graphical unification of fermion-to-qubit encodings","Fermion encodings: binary, tree, local—all ZX"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000748,"raw_usage":{"total_tokens":3404,"prompt_tokens":1089,"completion_tokens":2315,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":705,"completion_tokens_details":{"reasoning_tokens":2228}},"tokens_in":705,"tokens_out":2315,"duration_ms":16371,"temperature":1.0,"reasoning_tokens":2228,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:45:51.968683+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete ternary tree, compute its encoding matrix with Algorithm 1, and independently compute the matrix from the Pauli strings produced by the cited pairing scheme; any mismatch between the two matrices would falsify Theorem 2.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the scalable ZX-calculus with bold register wires, divide and gather nodes, and matrix arrows, the notation used for every diagram in the paper."},{"cited_title":"In Anca Muscholl, editor: Foundations of Software Science and Computation Structures , Lec- ture Notes in Computer Science, Springer, Berlin, Heidelberg, pp","cited_arxiv_id":null,"evidence_quote":"Establishes the phase-free normal form for interacting bialgebras, which is what reduces a phase-free diagram to a single matrix arrow."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the pushing-through-isometry technique that lets Jordan-Wigner operators be commuted past encoder diagrams to obtain encoded Pauli operators."}],"review_version":1}