{"id":"1950f24a-24c2-467f-a761-11e4d6b877f9","arxiv_id":"2509.03413","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A two-measurement syndrome extraction and teleportation-based recovery corrects single qubit insertion errors on gapped permutation-invariant codes.","lead":"A new quantum error correction protocol detects and corrects single insertion errors, which add an extra qubit at an unknown position, on a family of permutation-invariant codes. It uses two angular momentum measurements to produce a two-bit syndrome, then teleports the state back to a valid code.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Recovery protocol's logical-X maps the syndrome code to a different syndrome sector, so the teleportation cannot work as written","rationale":"The reader's weakest_assumption correctly flagged the recovery protocol, but the stress-test reveals a concrete inconsistency rather than just a missing proof: the explicitly defined X_L in Eq. (18) maps the measured syndrome sector to a different sector, so it cannot serve as the required logical-X for teleportation. This invalidates the central claim of correcting single insertion errors, since the j=N−1/2 case relies on the teleportation recovery. The syndrome-extraction part appears well-developed with detailed CG calculations, but the recovery step is load-bearing and currently broken as written. A REJECT (or at minimum major revision) is warranted because the presented protocol does not achieve the abstract's claimed correction; a fix would require redefining X_L or the CNOT to act within each syndrome code, with a proof of correctness.","tokens_in":18766,"tokens_out":10787,"duration_ms":107966,"concrete_test":"For the four-qubit gnu code (g=n=2,u=1,N=4), take the projected states |Ψ^{a,0}_{3/2}⟩ and |Ψ^{a,1}_{3/2}⟩ from Eq. (17). Apply X_L:m→−m to |Ψ^{a,0}_{3/2}⟩ and compute its squared overlap with the w=0 subspace span{|3/2,3/2⟩_p, |3/2,−1/2⟩_p} and with the w=1 subspace span{|3/2,−3/2⟩_p, |3/2,1/2⟩_p}. If the overlap with w=0 is 0 and with w=1 is 1, X_L does not preserve the syndrome code, so Eq. (19) is not a logical CNOT and the Fig. 3 teleportation fails. This analytic check uses only the paper's own definitions and can be done symbolically or numerically.","verdict_should_be":"REJECT","load_bearing_attack":"The recovery protocol hinges on the logical-X gate X_L in Eq. (18), defined by |j,m⟩_p → |j,−m⟩_p, being a logical Pauli-X within the syndrome-(j,w) code. This is not the case. For a state with j+m≡w (mod g), X_L sends m to −m, giving j+m' = j−m ≡ 2j−w (mod g). Thus X_L maps the w-sector to w' = 2j−w (mod g). For j=N+1/2, 2j=N+1=gnu+1≡1 (mod g), so w' = 1−w, swapping w=0 and w=1. For j=N−1/2, 2j=gnu−1≡−1 (mod g), so w' = −1−w, swapping w=0 and w=g−1. In no case is w'=w, so X_L maps the code subspace to an orthogonal syndrome subspace; it is a code-switching map, not a logical operator on the code. Consequently, in the CNOT of Eq. (19), when the ancilla |+_L⟩ provides control amplitude with odd m, the target leaves its codespace. The subsequent measurement of B in the logical-Z basis of the original w-sector then has non-zero probability of projecting onto the complement—or is undefined—so the Fig. 3 teleportation does not implement the claimed recovery. The assertion that a conditional logical-X is possible for odd g does not fix this (e.g., for the four-qubit code with g=2, X_L swaps w=0 and w=1).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a quantum error-correction protocol for single insertion errors on a family of permutation-invariant codes called gnu codes. After an insertion error increases the qubit count from N to N+1, the authors measure the total angular momentum J^2 and the magnetic quantum number J_z modulo the code gap g. They claim that these two measurements yield a two-bit syndrome (j,w) with j=N±1/2 and w restricted to two possible values, and that the post-measurement state is a valid quantum code on N+1 qubits. The central technical lemma (Lemma 1) asserts equality of the norms of the two projected logical codewords, which is proved in the appendix using Clebsch-Gordan identities, Vandermonde's identity, and binomial sums. The paper then describes a teleportation-based recovery protocol, using an ancilla in a gnu code, a logical controlled-NOT, and a conditional logical-X gate, intended to map the projected state back to a permutation-invariant code on the desired number of qubits.","tokens_in":19159,"tokens_out":7574,"duration_ms":82227,"significance":"If the protocol were correct, it would be the first general angular-momentum-based QEC protocol for single quantum insertion errors. The syndrome extraction stage is worked out in non-trivial detail: the post-insertion projections onto the symmetric and mixed-symmetry sectors are computed explicitly, and Lemma 1 is a substantive norm-equality result with a self-contained proof. The approach is parameter-free and uses only standard CG and binomial identities. However, the recovery stage as written is not sound. The logical-X gate defined in Eq. (18) does not preserve the syndrome sector of the projected code, so the logical-CNOT of Eq. (19) can move the target state out of its code subspace, and the subsequent logical-Z measurement on that subspace is not a valid measurement of the resulting state. Because the abstract's central claim ('we detail a QEC protocol that can correct single insertion errors') depends on this recovery procedure, the paper needs a major revision before it can be accepted.","major_comments":[{"comment":"The logical-X gate X_L: |j,m>_p -> |j,-m>_p is not a logical operator on the syndrome-(j,w) code. The code subspace is defined by j+m≡w (mod g). Under X_L, a state with quantum number m maps to m'=-m, so j+m' = 2j-w (mod g). For j=N+1/2, 2j = gnu+1 ≡ 1 (mod g), hence w' = 1-w and X_L swaps w=0 and w=1. For j=N-1/2, 2j ≡ -1 (mod g), hence w' = -1-w and X_L swaps w=0 and w=g-1. Thus X_L maps the code to an orthogonal syndrome sector, not to itself. Consequently, in the CNOT of Eq. (19), when the control qubit is in the logical |1> component, the target leaves its codespace. The subsequent measurement of register B in the logical-Z basis of the original sector is then undefined or has nonzero probability of projecting onto the complement, and the teleportation protocol of Fig. 3 does not implement the claimed recovery. The statement that a conditional logical-X is 'possible for odd code gap","section":"Recovery, Eq. (18) and Fig. 3"},{"comment":"The caption asserts that 'The logical controlled-NOT gate CAXB acts identically on |Ψ^{a,w}_j> for all a=0,...,N.' While this is plausible for the symmetric sector j=N+1/2, where the post-insertion state is independent of a, the mixed-symmetry codewords |x^{a,w}_{N-1/2}> defined in the appendix explicitly depend on a through the coefficients d_{a,p} and β_{k,l}. No proof is given that a fixed logical CNOT is well-defined across all insertion positions. If X_L is replaced by a code-preserving logical-X (as required by the previous comment), this point may become straightforward, but as written the claim is unsupported and essential to the recovery argument.","section":"Fig. 3 caption and Recovery paragraph"},{"comment":"The statement 'Such a unitary exists by the Knill-Laflamme QEC criterion' is not justified. The Knill-Laflamme condition gives a necessary and sufficient condition for a set of errors to be correctable by some recovery operation, but it does not by itself assert the existence of a unitary that maps one specific code space to another. If the authors intend to use a known result or construction, they should cite it; otherwise the claim should be removed or replaced with an explicit construction.","section":"Recovery section, second paragraph"}],"minor_comments":[{"comment":"Lemma 1 states 'Proof. Omitted for brevity.' but a proof is provided in the appendix. This is harmless, but the statement should read 'Proof: see Appendix' to avoid confusion.","section":"Main text, Lemma 1"},{"comment":"The protocol is described as yielding a 'two-bit syndrome' although w takes g possible values. The paper correctly notes that only two values of w have non-zero projection, but this is an important structural fact; it would help to state explicitly that the syndrome is effectively two bits because the two possible w values depend on j and the code gap.","section":"Eq. (14) and surrounding text"},{"comment":"The index notation in the norm computations is dense and occasionally ambiguous (e.g., the use of k, l, l' and the CG-coefficient subscripts). A short explanation of the summation ranges and the orthogonalization of the SCB would improve readability.","section":"Appendix, Eqs. (35)-(37) and (55)"},{"comment":"The normalization of the displayed states in Eq. (16) is not immediately obvious; a brief comment confirming that the states are normalized and orthogonal would be useful.","section":"Example 1, Eq. (16)"}],"recommendation":"major_revision","confidential_remarks":"The syndrome extraction part of the paper is substantial and appears technically sound. The main weakness is the recovery protocol: the logical-X defined in Eq. (18) is a code-switching map, not a logical operator on the syndrome code, and the teleportation protocol of Fig. 3 does not work as written. This is a fixable but non-local issue: the authors need to either find a code-preserving logical-X (or an alternative recovery) or rigorously show that a code-switching protocol can be adapted. I recommend major revision rather than rejection, because the syndrome extraction and the norm-equality lemma are valuable and the recovery can in principle be replaced."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here's my take on Bulled and Ouyang. The genuinely new thing is the two-stage angular-momentum syndrome extraction for insertion errors: measure J^2 then J_z mod g, and the claim that the two resulting bits (j,w) project the post-insertion state onto a well-defined codespace. The norm-equality lemma is proved in the appendix with standard CG and binomial identities, and the four-qubit example is concrete. That part looks right and is a real step toward a quantum analogue of insdel codes. The paper also honestly frames itself as building on Ouyang and Brennen's AM program, and the citations to that program and to [29] are appropriate.\n\nThe soft spot is the recovery protocol, and it's not minor. The logical-X gate in Eq. (18), |j,m> -> |j,-m>, does not preserve the syndrome sector. For a code with fixed j and w, applying X_L sends m to -m, so j+m becomes j-m, which is congruent to 2j-w mod g. Since 2j = gnu ± 1, you get w' = 1-w for j=N+1/2 and w' = -1-w for j=N-1/2. In no case is w'=w. So X_L maps the code to an orthogonal syndrome subspace; it is a code-switching map, not a logical operator. The CNOT of Eq. (19) then acts on the target by X_L^m, which sends a legitimate codeword into the wrong sector when m is odd, and the subsequent measurement in the original logical-Z basis will not reliably recover the state. The figure-3 caption asserts CAXB acts identically for all insertion positions without proof, and the assertion about odd g does not fix the arithmetic. This is a load-bearing flaw: the paper's central claim is correcting insertion errors, and the proposed recovery doesn't do that as written.\n\nThe syndrome-extraction half survives, and the idea of projecting onto an AM sector modulo g is worth keeping. But the recovery needs a different construction, or a revised definition of X_L that actually preserves w. As is, the protocol is incomplete.\n\nMy recommendation: send it to a serious referee. The mathematical core is nontrivial and mostly sound, and the flaw is specific and fixable in principle. But don't let it through as is; the recovery must be reworked or proven. If you read it, skip the discussion of GPGs and focus on Eqs. (18)-(19) and Fig. 3.","headline":"Syndrome extraction is a solid new piece; the recovery protocol's logical-X changes the syndrome sector, so the central correction claim is not supported as written.","tokens_in":19574,"tokens_out":4504,"would_cite":false,"duration_ms":41729,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P70","81P45","94B60"],"pacs":["03.67.Pp"],"model":"deepseek-v4-flash","headline":"The paper proposes a two-stage syndrome measurement of total angular momentum and its z-projection modulo the code gap to correct single insertion errors on gapped permutation-invariant codes, with teleportation-based recovery.","keywords":["quantum insertion errors","permutation-invariant codes","gnu codes","angular momentum","syndrome extraction","teleportation recovery","synchronisation errors","Clebsch-Gordan coefficients"],"falsifier":"Simulate the full recovery on the four-qubit code (g=n=2,u=1) for every insertion position a=0,...,4: prepare a logical state, insert a qubit, measure J^2 then J_z mod 2, run the proposed logical-CNOT teleportation with a |+_L> ancilla, and compare the final register with the original logical state after the known correction. If the logical CNOT's action depends on a, if the mixed-symmetry branch produces nonzero weight for a syndrome outside {0,g-1}, or if Lemma 1's equal-norm condition fails numerically, the central claim is false. This is a finite, directly checkable calculation.","tokens_in":18671,"feed_emoji":"⚛️","tokens_out":7260,"duration_ms":65749,"temperature":0.7,"pith_summary":"The paper claims the first general quantum error-correction protocol that corrects a single insertion error—an unknown extra qubit appearing among the encoded qubits—on gapped permutation-invariant (gnu) codes. The entire syndrome is two classical bits, obtained by measuring the total angular momentum J^2 and its z-projection modulo the code gap g. Those two global measurements project the corrupted state onto a new codespace, and the paper proves that the projected logical states form a legitimate orthonormal code. A teleportation circuit then returns the state to a permutation-invariant code on the desired number of qubits, using only operations implementable with geometric phase gates. If correct, this gives a general correction framework for a dimension-changing error class that previously had only bespoke examples.","feed_headline":"Two measurements correct single qubit insertions","feed_subtitle":"Measuring total spin and its z-projection yields a two-bit syndrome; teleportation restores the code.","key_machinery":"The machinery is angular-momentum coupling plus two projective measurements. The encoded state is a spin-N/2 object; an inserted qubit is a spin-1/2 object, so angular-momentum addition forces the whole system into total angular momentum j=N+1/2 (symmetric) or j=N-1/2 (mixed-symmetry). The projector P_j selects one of these sectors, and P^w_j selects states whose magnetic quantum number m satisfies j+m≡w mod g. Schur-Weyl duality provides the basis in which these projections dephase the state, and Clebsch-Gordan coefficients carry the position dependence. The equal-norm Lemma 1 is the identity that turns the projected states into orthonormal logical codewords, which is what allows a single t","core_discovery":"For a gnu code on N=gnu qubits, a single inserted qubit is treated as a spin-1/2 particle coupled to the logical state's spin N/2. Measuring J^2 then J_z modulo the code gap g projects the post-insertion state onto one of four syndromes—(j,w) with j=N+1/2, w=0,1 or j=N-1/2, w=0,g-1—and, crucially, Lemma 1 shows the two projected logical codewords for each syndrome have equal norm, so they genuinely encode a qubit. A logical-CNOT teleportation circuit, implemented with geometric phase gates, then maps that projected spin code back to a permutation-invariant code on the desired number of qubits. The four-qubit gnu code with g=n=2,u=1 is worked out explicitly.","pith_inferences":["The same J^2, J_z mod g syndrome is a plausible candidate for detecting and correcting deletion errors too, since a missing qubit also changes the total angular momentum sector; if that holds, it would give a unified insertion-deletion framework without relying on the classical Levenshtein equivalence.","The position-independence of the logical CNOT is the one step a numerical simulation should probe first: check the action of the logical CNOT on the projected states for every insertion position on a small gnu code; if it misbehaves, a Schur-transform-based recovery could replace the teleportation step.","Lemma 1's proof only needs the squared Clebsch-Gordan coefficients to be linear in the Dicke index, so the equal-norm property may extend to other permutation-invariant code families with affine weight profiles; verifying the norm-preserving condition for those families would test this.","The odd-gap requirement for the conditional logical-X gate suggests an open design question: either devise a modified recovery for even-gap gnu codes or prove that the two-measurement syndrome cannot support teleportation recovery there."],"forward_implications":["Single insertion errors on any gnu code with gap g are correctable with a two-bit syndrome, making decoding a lookup rather than a search.","When the first measurement gives j=N+1/2, the projected state already lives in the symmetric space on N+1 qubits, so the protocol can switch to a permutation-invariant code with better error-correction properties instead of merely undoing the insertion.","When it gives j=N-1/2, the teleportation step can map to either N or N+1 qubits, giving flexible code-length recovery.","Because every stage is expressible with geometric phase gates, the protocol does not require individual qubit addressability, which matters for photonic and bosonic hardware.","The four-qubit code example exhibits the general syndrome patterns, showing the framework is not vacuous at the smallest code size."],"supporting_citations":[{"why":"Defines the gnu code family and gives the four-qubit and Ruskai examples; the protocol's target states are gnu codes.","marker":"[1]"},{"why":"Supplies the four-qubit code used as the concrete example of the projected states for both syndrome branches.","marker":"[13]"},{"why":"Introduces the angular-momentum syndrome extraction approach and the teleportation-with-permutation-invariant-ancilla recovery that this protocol adapts.","marker":"[20]"},{"why":"Shows how to implement state preparation, measurements, and gates with geometric phase gates, which the protocol uses for implementation.","marker":"[23]"},{"why":"Provides the binomial-sum identities and gnu-code coefficient structure used in the proof of Lemma 1.","marker":"[29]"},{"why":"Supplies Schur-Weyl duality, which justifies projecting onto total-angular-momentum sectors in the Schur-Weyl basis.","marker":"[41]"},{"why":"Gives the theorem expressing Clebsch-Gordan coefficients as binomial-square roots, used throughout Lemma 1's norm computations.","marker":"[44]"},{"why":"Provides the recursion relation for Clebsch-Gordan coefficients used to simplify the mixed-symmetry norms.","marker":"[45]"},{"why":"Provides the teleportation circuit identity that the recovery circuit adapts.","marker":"[46]"}],"fun_headline_variants":["Measure twice, fix qubit insertions","Spin measurements fix inserted qubits","Two spin readings correct a single insertion","Angular momentum syndrome catches qubit insertions"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The recovery step assumes that the logical controlled-NOT gate treats the projected post-insertion state identically no matter where the extra qubit was inserted, and that a conditional logical-X gate can be built for odd code gaps; both are asserted without derivation, and if either is false the teleportation recovery fails.","fun_headline_variants_meta":{"raw":{"variants":["Measure twice, fix qubit insertions","Spin measurements fix inserted qubits","Two spin readings correct a single insertion","Angular momentum syndrome catches qubit insertions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000923,"raw_usage":{"total_tokens":3765,"prompt_tokens":688,"completion_tokens":3077,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":432,"completion_tokens_details":{"reasoning_tokens":3024}},"tokens_in":432,"tokens_out":3077,"duration_ms":22248,"temperature":1.0,"reasoning_tokens":3024,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T10:54:37.685727+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the full recovery on the four-qubit code (g=n=2,u=1) for every insertion position a=0,...,4: prepare a logical state, insert a qubit, measure J^2 then J_z mod 2, run the proposed logical-CNOT teleportation with a |+_L> ancilla, and compare the final register with the original logical state after the known correction. If the logical CNOT's action depends on a, if the mixed-symmetry branch produces nonzero weight for a syndrome outside {0,g-1}, or if Lemma 1's equal-norm condition fails numerically, the central claim is false. This is a finite, directly checkable calculation.","supporting_citations":[{"cited_title":"Permutation-invariant quantum coding for quantum deletion channels","cited_arxiv_id":"2102.02494","evidence_quote":"Defines the gnu code family and gives the four-qubit and Ruskai examples; the protocol's target states are gnu codes."},{"cited_title":"Nakayama and M","cited_arxiv_id":null,"evidence_quote":"Supplies the four-qubit code used as the concrete example of the projected states for both syndrome branches."},{"cited_title":"Necessary and sufficient condition for constructing a single qudit insertion/deletion code and its decoding algorithm","cited_arxiv_id":"2501.07027","evidence_quote":"Introduces the angular-momentum syndrome extraction approach and the teleportation-with-permutation-invariant-ancilla recovery that this protocol adapts."},{"cited_title":"Wang and P","cited_arxiv_id":null,"evidence_quote":"Shows how to implement state preparation, measurements, and gates with geometric phase gates, which the protocol uses for implementation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the binomial-sum identities and gnu-code coefficient structure used in the proof of Lemma 1."},{"cited_title":"Bacon, I","cited_arxiv_id":null,"evidence_quote":"Supplies Schur-Weyl duality, which justifies projecting onto total-angular-momentum sectors in the Schur-Weyl basis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the theorem expressing Clebsch-Gordan coefficients as binomial-square roots, used throughout Lemma 1's norm computations."},{"cited_title":"Clebsch-Gordan coefficients and the binomial distribution","cited_arxiv_id":"quant-ph/0112096","evidence_quote":"Provides the recursion relation for Clebsch-Gordan coefficients used to simplify the mixed-symmetry norms."}],"review_version":1}