{"id":"463dca8f-cd14-4888-b2c4-0c93f0eddf93","arxiv_id":"2412.07578","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every product-preserving ternary tree fermion-qubit mapping is equivalent to a linear encoding of the Fock basis, with an explicit matrix constructed from the tree.","lead":"Two families of fermion-to-qubit mappings, ternary tree transformations and linear Fock-basis encodings, are shown to coincide, up to relabelling and sign conventions, whenever the encoded vacuum is a product state. The paper supplies a shared notation and an explicit invertible binary matrix for every ternary tree, unifying separate toolkits used in quantum simulation.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 6.2 contains a sign error: with Γ_{2i+1}=-i bΓ_{2i+1} the vacuum stabilizer is -bΓ_{2i}bΓ_{2i+1}, so |0⟩^n is not the vacuum and m(T) is not a linear encoding as stated.","rationale":"Good-faith reading: the paper's main claim is plausible and the unified notation is valuable; the completeness argument would follow from Lemmas 5.9, 6.1 and 6.2. The reader correctly notes that Lemma 5.9's uniqueness classification is informal, but the more concrete breakdown is in Lemma 6.2's construction: a sign in Γ_{2i+1} flips every vacuum stabilizer. The n=1 instantiation shows the constructed mapping is affine but not linear, so as written the theorem does not prove existence of m(T). This is an internal inconsistency, not a disagreement with consensus. Since a one-character sign correction and a small proof amendment restore the argument, the appropriate outcome remains CONDITIONAL rather than ACCEPT or REJECT. The reader's weakest-assumption discussion did not identify this exact sign issue, hence partial agreement.","tokens_in":47610,"tokens_out":23765,"duration_ms":222454,"concrete_test":"Run the Lemma 6.2 construction for the unique one-vertex ternary tree (n=1): create eGT={X0,Y0,Z0}, let bΓ0=X0 and bΓ1=(-i)Y0; assemble m(T) as in the paper, Γ0=X0, Γ1=-i bΓ1=-Y0; then compute the vacuum stabilizer -iΓ0Γ1|0⟩ = -Z0|0⟩ = -|0⟩. If the result is -|0⟩ (as this algebra gives), Lemma 6.2's claim that |0_{m(T)}⟩=|0⟩^n fails. Equivalently, verify the symbolic identity -iΓ_{2i}Γ_{2i+1} = -bΓ_{2i}bΓ_{2i+1} against Eq. (63).","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 6.1, Lemma 6.2 defines bΓ_i=(-i)^{#y(eΓ_i)}eΓ_i, proves Eq. (63) that bΓ_{2i}bΓ_{2i+1}|0⟩^n=|0⟩^n, and then sets Γ_{2i}=bΓ_{2i}, Γ_{2i+1}=-i bΓ_{2i+1}. But for the vacuum stabilizer, -iΓ_{2i}Γ_{2i+1} = -i bΓ_{2i}(-i bΓ_{2i+1}) = -bΓ_{2i}bΓ_{2i+1}. Acting on |0⟩^n this gives -|0⟩^n, not +|0⟩^n. Hence |0⟩^n is a (-1)-eigenstate of each stabilizer and the actual vacuum of the constructed m(T) is not |0⟩^n. The n=1 case is explicit: eGT={X0,Y0,Z0}, bΓ0=X0, bΓ1=-iY0, so Γ0=X0, Γ1=-Y0 and -iΓ0Γ1=-Z0; the vacuum is |1⟩, and the Fock states are |f⟩→|f⊕1⟩, an affine but not linear encoding. This contradicts Lemma 6.2 property 2 and therefore invalidates the construction of m(T) used in Theorem 2. The error is readily repaired by taking Γ_{2i+1}=+i bΓ_{2i+1} (the sign change is an allowed equivalence), but as written the central constructive proof is internally inconsistent.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a unified framework for ancilla-free fermion-qubit mappings, connecting the operator-based definition (ordered pairs of anticommuting Pauli strings) with the state-based definition (encodings of the Fock basis). It introduces an equivalence relation on mappings, defines classical, affine, and linear encodings, and then gives a refined definition of ternary tree transformations. The central claim is Theorem 2: for every ternary tree T there is a unique T-based mapping m(T) that linearly encodes the Fock basis, and every product-preserving ternary tree transformation is equivalent, under the paper's equivalence relation, to one of these m(T). The paper also identifies, for the complete ternary tree, the resulting linear encoding with the pruned Sierpinski tree transform.","tokens_in":47972,"tokens_out":5482,"duration_ms":53726,"significance":"If the main theorem is correct, the paper establishes a genuine conceptual equivalence between two classes of fermion-qubit mappings that have usually been treated separately: product-preserving ternary tree transformations are not an independent class but are contained, up to the paper's equivalence, in linear encodings of the Fock basis. This is a useful and non-obvious result, and the paper gives substantial supporting apparatus: a unified notational framework, a taxonomy of equivalence templates, a formula for the binary matrix G_T that defines the encoding m(T), and a concrete identification with the pruned Sierpinski transform. The paper is largely self-contained and the proofs are detailed, which is a strength. The sign defect discussed below is local and repairable, but it affects the central construction as written.","major_comments":[{"comment":"The construction of m(T) contains a sign error that invalidates the claimed vacuum and therefore the claimed linearity. The authors define bΓ_i = (-i)^{#y(eΓ_i)} eΓ_i, prove Eq. (63) that bΓ_{2i}bΓ_{2i+1}|0>^n = |0>^n, and then set Γ_{2i}=bΓ_{2i} and Γ_{2i+1}=-i bΓ_{2i+1}. But then -iΓ_{2i}Γ_{2i+1} = -bΓ_{2i}bΓ_{2i+1}, so the vacuum stabilizer acts on |0>^n as -|0>^n. Thus |0>^n is not the vacuum state of the constructed mapping. The n=1 case makes the failure explicit: eΓ_0=X0, eΓ_1=Y0 gives bΓ_0=X0, bΓ_1=-iY0, hence Γ_0=X0 and Γ_1=-Y0, so -iΓ_0Γ_1=-Z0 and the vacuum is |1>, not |0>. Consequently the Fock states are |f> -> |f⊕1>, an affine but not linear encoding. This contradicts Lemma 6.2 property 2 and the existence half of Theorem 2(a). Replacing Γ_{2i+1}=-i bΓ_{2i+1} with +i bΓ_{2i+1} repairs the construction, and this sign change is an allowed equivalence under Definition 3.4, but as written the central constructive proof is internally inconsistent.","section":"6.1, Lemma 6.2, Eq. (63) and the paragraph after Eq. (92)"},{"comment":"The classification of all possible operator pairings that preserve a product vacuum is load-bearing for the completeness claim in Theorem 2(b), but the proof is not rigorous. The text rules out alternative pairing structures with the sentence beginning 'But because there are only three mutually anticommuting single-qubit Pauli matrices', asserting that elements from distinct pairs would have to anticommute on a child vertex and that this would make it impossible for the product state to be an eigenstate of both products. This is a plausibility argument, not a formal proof; a full case analysis is needed to exclude exotic pairing patterns. Without a rigorous Lemma 5.9, the uniqueness assertion in Theorem 2(a) and the completeness assertion in Theorem 2(b) are not fully supported.","section":"5.1, Lemma 5.9, proof of part (a), paragraph after Eq. (50)"}],"minor_comments":[{"comment":"The final sentence of the abstract states that 'every ternary tree transformation' is equivalent to a linear encoding, but the theorem and body of the paper only claim this for product-preserving ternary tree transformations; the qualifier should be added to the abstract as well.","section":"Abstract (full text)"},{"comment":"In the proof of Lemma 6.2, the sentence 'Theorem 3 proved that every classical encoding is affine' appears to reference the wrong result; Theorem 1 in Section 4.1, or Corollary 4.7, is the relevant statement that classical encodings with Pauli representations are affine.","section":"6.1, Lemma 6.2"},{"comment":"The caption mentions 'the ternary tree transformation mTT' without defining it; either define this notation or rephrase the caption to refer to the complete ternary tree transformation.","section":"Introduction, Figure 1 caption"},{"comment":"There is a duplicated word in 'The link is via a unique unique unitary operator'; this should be corrected.","section":"Section 2, text before Eq. (27)"}],"recommendation":"major_revision","confidential_remarks":"The sign error in Lemma 6.2 is easily repairable, and the paper's thesis is plausible, so I would not reject the manuscript. However, the authors should not merely patch the prefactor in Lemma 6.2; they should also tighten the classification argument in Lemma 5.9, since the completeness of Theorem 2 depends on it. The current version is not acceptable as is, but the route to a correct and publishable paper is clear."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe paper does something real: it gives a common language for operator-first (ternary tree) and state-first (linear encoding) fermion-qubit mappings, proves that classical encodings are affine and hence equivalent to linear ones, and constructs for each ternary tree T an explicit binary matrix G_T of a T-based linear encoding. If the construction holds, this closes a gap several of us have been wondering about and gives practitioners a dictionary between tree diagrams and CNOT circuits. That alone is worth a serious look. Section 4's treatment of affine encodings and the equivalence notion in Section 3 are clean and useful. I also appreciate that Lemma 6.2 is an explicit constructive algorithm, not an existence argument.\n\nBut there is a concrete problem in the construction of m(T). Lemma 6.2 defines bΓ_i = (-i)^{#y(eΓ_i)} eΓ_i, proves bΓ_{2i} bΓ_{2i+1}|0>^n = |0>^n, then sets Γ_{2i}=bΓ_{2i}, Γ_{2i+1}=-i bΓ_{2i+1}. The vacuum stabilizer is -iΓ_{2i}Γ_{2i+1} = -bΓ_{2i}bΓ_{2i+1}, so |0>^n is a (-1)-eigenstate, not the vacuum. The n=1 case is explicit: Γ0=X, Γ1=-Y, stabilizer -Z, vacuum |1>, and the Fock encoding is affine (|f> maps to |f⊕1>) rather than linear. This contradicts property 2 of Lemma 6.2 and therefore Theorem 2 as written. The repair is easy — take Γ_{2i+1}=+i bΓ_{2i+1} — and it is a local sign convention, but as submitted the central constructive proof is internally inconsistent.\n\nOther soft spots are minor by comparison. Lemma 5.9's classification has an informal step ('only three Pauli matrices') that I found plausible but not fully rigorous. The long induction in Lemma 6.2 is not machine-checked, and it carries real bookkeeping weight; after the sign error, I would want that induction checked more carefully before trusting it. The claimed equality with the pruned Sierpinski transform is demonstrated by figures and recursive pattern rather than a fully general proof. The title also overclaims by omitting 'product-preserving', though the abstract is accurate. Self-citations supply the Sierpinski identification, but the main equivalence is derived from definitions within the paper.\n\nWho is this for? Someone working on fermion-qubit mapping design or trying to reconcile tree-based and matrix-based encodings will get genuine value. It deserves a serious referee, but only after the sign fix and a tightening of Lemma 5.9. I would send it out.","headline":"A genuinely useful unification of ternary tree and linear-encoding fermion-qubit mappings, but the central constructive lemma has a sign error that must be fixed before the theorem as stated is reliable.","tokens_in":48462,"tokens_out":2767,"would_cite":true,"duration_ms":26118,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Ac"],"model":"deepseek-v4-flash","headline":"This paper proves that every product-preserving ternary tree transformation is equivalent to a linear encoding of the Fock basis, with an explicit invertible binary matrix $G_T$ for each ternary tree $T$.","keywords":["fermion-qubit mapping","ternary tree transformation","linear encoding","Fock basis","Majorana operators","Clifford group","quantum simulation","Sierpinski tree transform"],"falsifier":"Enumerate, for a small ternary tree such as the five-vertex tree in the paper's Example 5.4, every $T$-based mapping whose vacuum is a product state, and check whether each one lies in the equivalence class of $m(T')$ for some tree $T'$ obtained by local Pauli relabellings; any product-preserving $T$-based mapping found outside all such classes would refute Theorem 2.","tokens_in":47422,"feed_emoji":"🌳","tokens_out":8060,"duration_ms":64166,"temperature":0.7,"pith_summary":"This paper claims that two seemingly different design styles for fermion-to-qubit mappings are actually one class: every product-preserving ternary tree transformation is equivalent to a linear encoding of the Fock basis. In a linear encoding, each fermionic occupation vector $f$ is stored as a computational basis state $\\lvert Gf\\rangle$ for an invertible binary matrix $G$; the Jordan-Wigner and Bravyi-Kitaev transformations are examples. Product-preserving means the encoded vacuum state is a tensor product of single-qubit states. The paper develops a unified operator-based and state-based notation for fermion-qubit mappings, defines an equivalence relation that factors out labelling and sign choices, and proves that for each ternary tree $T$ there is a unique $T$-based mapping $m(T)$ that is also a linear encoding. If correct, this collapses two separate families of mappings into one and connects tree-based minimal-weight constructions to the searchable space of binary matrices.","feed_headline":"Every tree-based fermion mapping is secretly a linear encoding","feed_subtitle":"A unified notation shows tree Pauli strings and Fock-basis matrices are the same mapping class; the Sierpinski case is the complete tree.","key_machinery":"The central object is the $T$-based mapping $m(T)$, defined as a fermion-qubit mapping whose $2n$ Majorana-representing Pauli operators are signed elements of the maximally anticommuting set $\\tilde G_T$ obtained from the root-to-leaf paths of the ternary tree $T$. The argument is carried by three lemmas: Lemma 5.9 shows that for any chosen product stabiliser vacuum state, the pairing of the tree Pauli operators that preserves that vacuum is unique up to fermionic relabelling and pair braids; Lemma 6.2 constructs the unique pairing that is also a classical encoding, using a vertical path-ordering scheme with Y-branch inversions and phase factors $(-i)^{\\#_y}$; Lemma 6.3 gives the matrix formula $(G_T)_{ij}=1$ exactly when $\\Gamma_{2j}$ acts on qubit $i$ by $X$ or $Y$. This machinery turns a tree graph directly into an invertible binary matrix.","core_discovery":"On its own terms, the central discovery is Theorem 2: for every $n$-vertex ternary tree $T$ there is a unique $T$-based fermion-qubit mapping $m(T)$ that both is built from the anticommuting Pauli strings associated with the root-to-leaf paths of $T$ and linearly encodes the Fock basis, meaning $\\lvert f_{m(T)}\\rangle = \\lvert G_T f\\rangle$ for an explicit invertible binary matrix $G_T$; and every $n$-mode product-preserving ternary tree transformation is equivalent to some $m(T)$ under the paper's equivalence relation of qubit relabelling, local Pauli basis changes, Pauli pair braids, sign changes, and fermionic relabelling. The proof constructs the Clifford operator $C_T$ by ordering the $2n+1$ tree paths vertically, inverting the order after Y-branches, and defining phase-corrected operators $\\hat\\Gamma_i = (-i)^{\\#_y(\\tilde\\Gamma_i)}\\tilde\\Gamma_i$; it then sets $\\Gamma_{2i}=\\hat\\Gamma_{2i}$ and $\\Gamma_{2i+1}=-i\\hat\\Gamma_{2i+1}$. As a concrete payoff, applying the construction to the complete ternary tree recovers the pruned Sierpinski tree transform, so the two existing literatures describe the same object.","pith_inferences":["If Theorem 2 holds, then any optimisation or hardware-oriented search conducted over linear encodings has already implicitly searched the space of product-preserving ternary tree transformations; the converse is not automatic for product-breaking mappings, which remain outside the equivalence.","The explicit matrix formula suggests that cost measures of tree-based mappings, such as Pauli weight or CNOT count, can be stated as functions of the matrix $G_T$ alone, potentially enabling matrix-based optimisation heuristics.","A natural next step is to characterise the image of the map $T \\mapsto G_T$, i.e., to determine which invertible binary matrices arise from ternary trees; the template equivalence suggests this image forms a finite catalogue for each $n$.","The equivalence also implies that the Bonsai and Treespilation search heuristics could in principle be re-expressed as searches over binary matrices with tree-compatible update sets, although the paper does not explicitly make this algorithmic translation."],"forward_implications":["For every $n$-vertex ternary tree $T$, there is exactly one $T$-based mapping that is also a linear encoding of the Fock basis, so product-preserving ternary tree transformations no longer need a separate operator-based treatment.","Any product-preserving ternary tree transformation can be represented by an invertible binary matrix $G_T$, with an explicit entry formula, so tree-based mappings inherit the update, parity, and flip rules of linear encodings.","The pruned Sierpinski tree transform is the special case $m(T)$ for the complete ternary tree, unifying two independently discovered minimal-weight constructions.","Computational searches over linear encodings already cover product-preserving ternary tree transformations, and a linear encoding can be checked for whether it is a ternary tree transformation using the paper's characterisation.","The template equivalence relation groups product-preserving tree-based mappings into classes that differ only by labelling and sign choices, so optimisation over these mappings can be performed on equivalence classes rather than individual Pauli strings."],"supporting_citations":[{"why":"Supplies the original ternary tree transformation construction and the proof that complete ternary trees achieve minimal average Pauli weight, which motivates the class studied here.","marker":"[5]"},{"why":"Extends ternary tree transformations and defines a Pauli pairing whose vacuum state is $\\lvert 0\\rangle^{\\otimes n}$, which Lemma 5.9 generalises to arbitrary product vacuum states.","marker":"[6]"},{"why":"Provides the Sierpinski data structure and Clifford implementation of $\\lvert f\\rangle \\mapsto \\lvert Gf\\rangle$ that the proof of Lemma 6.2 explicitly builds on for constructing $m(T)$.","marker":"[21]"},{"why":"Defines the generalised update, parity, flip, and remainder sets used in Definition 4.5 and Corollary B.7 to describe linear encodings.","marker":"[25]"},{"why":"Describes the Bonsai algorithm's pairing prescription for product-preserving ternary tree mappings, which Lemma 5.9 extends to any product vacuum state.","marker":"[27]"},{"why":"Presents Treespilation, a computational search over a subset of ternary tree transformations, referenced as a related prescription in Examples 5.11 and 5.12.","marker":"[28]"},{"why":"Introduces the pruned Sierpinski fermion-to-qubit transform, which the paper recovers as $m(T)$ for the complete ternary tree in Section 6.3.","marker":"[35]"},{"why":"Provides the Clifford group and stabiliser tableau background used to identify the Clifford operator $C_T$ and the affine form of classical encodings.","marker":"[33]"}],"fun_headline_variants":["Tree-based fermion mappings equal linear Fock encodings","Ternary tree transforms are linear encodings in disguise","Unified notation shows trees and matrices map identically","Complete ternary tree yields the Sierpinski transform","Every product-preserving tree map is a linear encoding"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The completeness half of Theorem 2 rests on Lemma 5.9's classification that every pairing of tree Pauli operators whose vacuum is a product state must have the form given in Equation 49, and the proof's argument against alternative pairing structures is informal, based on there being only three Pauli matrices per qubit.","fun_headline_variants_meta":{"raw":{"variants":["Tree-based fermion mappings equal linear Fock encodings","Ternary tree transforms are linear encodings in disguise","Unified notation shows trees and matrices map identically","Complete ternary tree yields the Sierpinski transform","Every product-preserving tree map is a linear encoding"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000225,"raw_usage":{"total_tokens":1478,"prompt_tokens":972,"completion_tokens":506,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":588,"completion_tokens_details":{"reasoning_tokens":439}},"tokens_in":588,"tokens_out":506,"duration_ms":5163,"temperature":1.0,"reasoning_tokens":439,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:42:47.412010+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Enumerate, for a small ternary tree such as the five-vertex tree in the paper's Example 5.4, every $T$-based mapping whose vacuum is a product state, and check whether each one lies in the equivalence class of $m(T')$ for some tree $T'$ obtained by local Pauli relabellings; any product-preserving $T$-based mapping found outside all such classes would refute Theorem 2.","supporting_citations":[{"cited_title":"Optimal fermion-to-qubit mapping via ternary trees with applications to reduced quantum states learning","cited_arxiv_id":null,"evidence_quote":"Supplies the original ternary tree transformation construction and the proof that complete ternary trees achieve minimal average Pauli weight, which motivates the class studied here."},{"cited_title":"Clifford Algebras, Spin Groups and Qubit Trees","cited_arxiv_id":null,"evidence_quote":"Extends ternary tree transformations and defines a Pauli pairing whose vacuum state is $\\lvert 0\\rangle^{\\otimes n}$, which Lemma 5.9 generalises to arbitrary product vacuum states."},{"cited_title":"A Sierpinski Triangle Data Structure for Efficient Array Value Update and Prefix Sum Calculation","cited_arxiv_id":"2403.03990","evidence_quote":"Provides the Sierpinski data structure and Clifford implementation of $\\lvert f\\rangle \\mapsto \\lvert Gf\\rangle$ that the proof of Lemma 6.2 explicitly builds on for constructing $m(T)$."},{"cited_title":"Fermion-to-qubit mappings with varying resource requirements for quantum simulation","cited_arxiv_id":null,"evidence_quote":"Defines the generalised update, parity, flip, and remainder sets used in Definition 4.5 and Corollary B.7 to describe linear encodings."},{"cited_title":"Bonsai Algorithm: Grow Your Own Fermion-to-Qubit Mappings","cited_arxiv_id":null,"evidence_quote":"Describes the Bonsai algorithm's pairing prescription for product-preserving ternary tree mappings, which Lemma 5.9 extends to any product vacuum state."},{"cited_title":"Treespilation: Architecture- and State-Optimised Fermion- to-Qubit Mappings, 2024","cited_arxiv_id":null,"evidence_quote":"Presents Treespilation, a computational search over a subset of ternary tree transformations, referenced as a related prescription in Examples 5.11 and 5.12."},{"cited_title":"Whit- field","cited_arxiv_id":null,"evidence_quote":"Introduces the pruned Sierpinski fermion-to-qubit transform, which the paper recovers as $m(T)$ for the complete ternary tree in Section 6.3."},{"cited_title":"The Clifford group, stabilizer states, and linear and quadratic operations over GF(2)","cited_arxiv_id":"quant-ph/0304125","evidence_quote":"Provides the Clifford group and stabiliser tableau background used to identify the Clifford operator $C_T$ and the affine form of classical encodings."}],"review_version":1}