{"id":"cfbc7981-d2cb-4df2-84fd-54366c320415","arxiv_id":"1908.00705","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors claim protocols for 1-to-2 and 1-to-3 asymmetric universal quantum clone multicast over quantum networks, but the 1-to-2 protocol has a critical error and works only for special measurement outcomes.","lead":"This paper proposes quantum network protocols that multicast asymmetric universal quantum clones, approximate copies with different fidelities, from one sender to two or three receivers. It extends earlier symmetric-clone network coding by Owari et al. and uses Kobayashi et al.'s quantum network coding as a subroutine.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The swap criticism does not land; the real flaw is the reciprocal β normalization in Eq. (28), which breaks Eq. (29) and the 1-to-2 claim.","rationale":"The Reader's identified weakest assumption is that Γ^{(r)} in Step 7 cannot move the d-level states |r−1⟩ and |r⟩ into the two-dimensional systems G and H. This does not survive inspection: Step 6 precedes Step 7 and applies V^{(r)} and Δ^{(r)}. For the r-branch, Δ^{(r)} uses X^{−(r−1)} on both E and F, sending |r−1⟩→|0⟩ and |r⟩→|1⟩, so the branch on which Γ acts (C=D=r) is always in the |0⟩,|1⟩ subspace. For j≠r branches, Γ is controlled off because C,D≠r, so no swap of |j⟩ or |r⟩ is attempted. A direct check for arbitrary r confirms Eq. (34) follows. Thus the swap objection is not the load-bearing problem.\n\nThe actual load-bearing flaw is a coefficient error in Eq. (28). The raw post-measurement branch from the stated isometry (2) has coefficients a/√d and b/√d; after the r-dependent unitary Υ^{(r)}, the surviving single-system coefficients are √(a^2+b^2)/√d and (a+b)/√d times the input amplitudes. The printed β_j and β_r are the reciprocals of these values. This is not a matter of convention: with the printed coefficients, the decomposition in Eq. (27) is not a Kraus decomposition of the 1→2 cloning channel, since ∑_r L_r† L_r is not the identity. In the symmetric qubit case it gives 7.5 times the identity, so the purported final ensemble is not even trace-preserving. Eq. (29), which is the bridge from the protocol to the claimed optimal fidelities, therefore fails.\n\nThe conclusion in Section V asserts both the 1-to-2 and 1-to-3 multicast results. Because the 1-to-2 proof is not established, the central claim of the abstract is unsupported as printed. This is an internal derivation error, not a disagreement with consensus, and it is specific enough to be settled by the symbolic check above.","tokens_in":23053,"tokens_out":29729,"duration_ms":289319,"concrete_test":"Recompute Eq. (29) symbolically for d=2 with a=b=1/√3 and |ψ⟩=|0⟩: substitute Eq. (27) with β from Eq. (28) into the left side, and compare with Tr_M[U_{1→2}|0⟩⟨0|U_{1→2}^†]. The left side has trace 7.5 and the right side trace 1, so the equivalence fails. Then repeat with corrected β_j = α_j√(1−2ab/d)/√d and β_r = α_r(a+b)/√d; the two sides should match exactly, confirming the location of the error.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III B's proof depends on the decomposition (27) and the coefficients (28). Tracing the measurement-outcome branch from U_{1→2} of Eq. (2) gives K_r|ψ⟩ = d^{-1/2} ∑_j α_j(a|jr⟩+b|rj⟩). Applying Υ^{(r)} of Eqs. (18)–(19) and discarding B turns this into d^{-1/2}[∑_{j≠r} α_j√(a^2+b^2)|j⟩ + α_r(a+b)|r⟩]. Therefore the coefficients in Eq. (30) must be β_j = α_j√(1−2ab/d)/√d and β_r = α_r(a+b)/√d. Eq. (28) states the reciprocals of these expressions. With the printed β, the operators L_r defined by Eq. (27) satisfy ∑_r L_r† L_r = [d(d−1)/(a^2+b^2) + d/(a+b)^2] I, which equals I only in special cases; for symmetric a=b=1/√3, d=2, the scalar is 7.5. Hence Eq. (29), ∑_r |Ψ_2^{(r)}⟩⟨Ψ_2^{(r)}| = ε_{1→2}(|ψ⟩⟨ψ|), is false, and the final state (39) is not the optimal asymmetric clone channel. This invalidates the 1-to-2 multicast claim in Section V.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes quantum multicast network coding protocols that distribute optimal asymmetric universal quantum clones (1→2 and 1→3) from one source node to two or three target nodes. The protocols combine Cerf's asymmetric cloning isometry, Kobayashi et al.'s quantum multicast network coding for GHZ-type states, local compression via controlled operations, teleportation, and a constant amount of preshared entanglement. The main claimed result is that, under the stated assumptions on the underlying classical network code, the protocol multicasts optimal asymmetric clones of a q^r-dimensional state while consuming entanglement that does not scale with q.","tokens_in":23304,"tokens_out":17273,"duration_ms":165126,"significance":"If the central claim held, this would be a meaningful extension of Owari et al.'s symmetric-cloning multicast protocols to asymmetric cloning and would give a constructive, constant-overhead quantum network-coding scheme for this task. The 1→3 protocol appears more robust and I did not find a comparable algebraic error there. The paper also benefits from grounding the protocol in the independently established optimal asymmetric cloning results, so there is no circularity in the performance check. However, the 1→2 proof contains a load-bearing normalization error in Eq. (28) that invalidates the central claim as written; the error is local and appears fixable.","major_comments":[{"comment":"The coefficients β_j and β_r printed in Eq. (28) are reciprocals of the values required by the actual state after measuring M in Step 2, and consequently Eq. (29) is false as stated. Directly tracing the outcome branch from Eq. (2) gives the unnormalized state d^{-1/2}[(a+b)α_r|rr> + Σ_{j≠r} α_j(a|jr> + b|rj>)]. Since cosη = a/c and sinη = b/c with c² = 1 - 2ab/d, this branch is β_r|rr> + Σ_{j≠r} β_j(cosη|jr> + sinη|rj>) with β_r = α_r(a+b)/√d and β_j = α_j c/√d, not the expressions in Eq. (28). With the printed β values, the trace of the left-hand side of Eq. (29) on a normalized input is d/(a+b)² + d(d-1)/c²; for example, for d=2 and a=b=1/√3 this trace is 9/2 instead of 1, so the channel is not trace-preserving and Eq. (29) cannot hold. This error propagates through Eq. (30) and the final state (39), so the 1→2 multicast claim in Section V is unsupported as printed. Replacing Eq. (28) with the correct coefficients above is a local fix and restores the intended argument.","section":"III B, Eq. (28)"}],"minor_comments":[{"comment":"The description of the preshared entanglement says \"cosη|0>_E|1>_E + sinη|1>_E|0>_F is shared between E and F\"; the first term should be |0>_E|1>_F, matching Eq. (31).","section":"III A"},{"comment":"The third fidelity in Eq. (10) is labeled F_A but should be F_C, since it concerns the output system C.","section":"II A, Eq. (10)"},{"comment":"The transition from Eq. (34) to Eq. (35) assumes that quantum teleportation delivers the state on H to T1 without a Pauli byproduct. The protocol should state explicitly that the standard teleportation corrections are applied; otherwise the derivation skips a necessary step. This is easily fixed given the assumption of free classical communication.","section":"III A, Protocol 2, Steps 8-10"},{"comment":"There are many typographical errors, including \"procol\" (Section III heading), \"faor\" (Section IV B), \"differs\" in the introduction, and inconsistent comma usage in Protocol 3 Step 2. The duplicate references [25]/[37], [26]/[38], and [30]/[39] should also be merged.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The main problem is a single normalization error in Eq. (28); it is serious because it breaks the central 1→2 proof, but it appears to be a local, fixable typo rather than a conceptual obstruction. The partial-swap step criticized in the reader's report is actually consistent: after Step 6 the r-branch lies in the |0>,|1> subspace of E and F, while the j≠r branches are untouched because of the |r> + b|rj>) + α_r(a+b)|rr>]. Matching Eq. (27) forces β_j = α_j √(1−2ab/d)/√d and β_r = α_r(a+b)/√d. The printed Eq. (28) gives the reciprocals. Consequently Eq. (29), the decomposition of the cloning channel, is false as written; for a=b=1/√3, d=2 the printed ensemble is off by a factor of 7.5. The final state (39) is then not the desired clone. This is not an inessential line: it is the proof that the protocol implements the cloner. The good news is that the protocol operations themselves do not depend on β, so replacing the printed β with the correct coefficients appears to make the whole 1-to-2 argument go through.\n\nThe swap criticism in the report is, I think, wrong. By Step 7 the E/F systems in the r-branch have been reduced to |0>/|1> by V^(r) and Δ^(r), so Γ^(r)'s swap acting only on the two-dimensional subspace is sufficient. Eq. (34) is consistent.\n\nWho this is for: quantum information people working on network coding and cloning. The 1-to-3 construction is worth a careful read; the 1-to-2 section needs a corrected proof before it should be cited. I would send this to referees, because a concrete, fixable coefficient error is exactly what review is for, but I would not accept the current version as is.","headline":"The 1-to-3 protocol looks genuinely sound, but the printed 1-to-2 proof has a coefficient error in Eq. (28) that breaks the main claim; the swap criticism in the report does not survive contact with the paper.","tokens_in":23835,"tokens_out":8827,"would_cite":false,"duration_ms":84014,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.-a","03.67.Hk"],"model":"deepseek-v4-flash","headline":"This paper constructs quantum network coding protocols that multicast 1→2 and 1→3 optimal asymmetric universal clones of a $q^r$-dimensional state, with constant shared entanglement.","keywords":["quantum network coding","multicast protocol","universal quantum cloning","asymmetric cloning","optimal fidelity","GHZ state","entanglement resource","LOCC"],"falsifier":"Apply the protocol with $q=3$ and measurement outcome $r=3$, and track the component $\\cos\\eta\\,|2\\rangle_E|3\\rangle_F$ through Step 7: since $\\Gamma^{(3)}$ is the identity whenever the control register is not in $|3\\rangle$, this component stays in $EF$ and never moves to $GH$, so Eq. (34) fails, and the final reduced state on $EF$ would show a fidelity below the predicted optimal asymmetric bounds.","tokens_in":22816,"feed_emoji":"🕸️","tokens_out":9544,"duration_ms":81891,"temperature":0.7,"pith_summary":"This paper constructs quantum network coding protocols that multicast approximate copies—optimal asymmetric universal clones—of an unknown $q^r$-dimensional quantum state from one source to two or three target nodes. Because no-cloning forbids perfect quantum multicast, the right goal is to deliver the best physically allowed imperfect copies, and the paper shows this can be done over a quantum network whose edges carry only $q$-dimensional systems, provided a classical linear multicast code of rate $r$ exists and classical communication is free. The target nodes need only a small amount of shared entanglement, at most 2 ebits for two targets and $2+4\\log_2 3$ ebits for three, an amount independent of $q$ and of the input dimension. If the construction is correct, asymmetric cloning—where receivers may be given clones of different quality—becomes a network-layer service rather than a source-local operation.","feed_headline":"Quantum networks can multicast asymmetric optimal clones","feed_subtitle":"A protocol distributes 1-to-2 and 1-to-3 optimal clones using entanglement that does not grow with dimension.","key_machinery":"The load-bearing object is a compression–reconstruction pair of unitaries that fits a branch state carrying the clone information into a $q$-dimensional transmission. On the source side, after the cloning isometry and ancilla measurement, the unitary $\\Upsilon^{(r)}$ rotates the conditional state into a single $d$-dimensional register; on the receiver side, the partial swaps $\\Gamma^{(r)}$ together with $V^{(r)}$, $\\Delta^{(r)}$, and $\\Lambda^{(r)}$ redistribute the $r$-dependent components between the $d$-dimensional GHZ-shared systems and small two-dimensional systems, so the clone branches are rebuilt locally. The pre-shared entanglement between targets supplies the two-level routing degrees of freedom that the transmitted $d$-dimensional state cannot carry.","core_discovery":"The central claim is that optimal asymmetric universal quantum cloning can be implemented as a distributed network protocol: the source applies the 1→2 (or 1→3) asymmetric cloning isometry, measures the ancilla, and compresses the conditional branch state into a single $d=q^r$-dimensional register; this register is multicast to all targets using a previously established quantum multicast subroutine that yields a shared GHZ-type state; and receivers then reconstruct their individual clones by local unitaries and measurements, using shared entangled pairs and the classical measurement outcome from the source. The protocol is claimed to end with the two (three) target nodes holding exactly the reduced states of an optimal asymmetric UQCM, with fidelities matching the optimal formulas, and with entanglement cost bounded by a constant independent of $q$. The 1→3 case is handled by a two-branch argument depending on whether the two ancilla measurement outcomes are equal or distinct.","pith_inferences":["If the partial-swap reconstruction for general $r$ is repaired, for example by extending $\\Gamma^{(r)}$ to move arbitrary $|r-1\\rangle,|r\\rangle$ pairs into the two-dimensional registers, the same scheme could plausibly extend to $1\\to n$ clone multicasts for $n\\ge 4$, since the $n=2,3$ cases need only $O(1)$ extra local dimensions.","The protocol effectively realizes a flagged quantum state split: the ancilla outcome $r$ selects which computational branch is amplified, and the small shared entanglement carries only this flag; this suggests a general template in which any optimal cloning machine with a discrete flag can be multicast with constant entanglement.","A concrete numerical test of the 1-to-2 protocol at $q=3$ would be to compare the two output reduced fidelities to $F_A=1-b^2(d-1)/d$ and $F_B=1-a^2(d-1)/d$; a mismatch for $r\\notin\\{1,2\\}$ would locate exactly where the branch reconstruction fails."],"forward_implications":["If the claim holds, any quantum network with a solvable rate-$r$ classical linear multicast code on an acyclic orientation can multicast asymmetric clones, making clone-fidelity asymmetry a tunable network feature.","The entanglement overhead is constant in $q$, so the cost of multicast cloning does not grow with the dimension of the cloned state; only the classical coding rate $r$ grows.","For sufficiently large $q$, the classical-code existence condition is equivalent to each target's min-cut from the source being at least $r$, so the protocol applies from network-topology data alone.","Setting the cloning parameters equal ($a=b$, or $\\alpha=\\beta=\\gamma$) recovers symmetric optimal clone multicast as the symmetric limit of the same protocol.","The 1-to-3 protocol explicitly branches on whether the two ancilla outcomes coincide, so the same transmitted $d$-dimensional state is decoded by two different reconstruction circuits."],"supporting_citations":[{"why":"Provides the quantum multicast subroutine used to distribute the compressed state as a GHZ-type state to all targets.","marker":"[8]"},{"why":"Defines the 1→2 optimal asymmetric universal cloning isometry and the optimal fidelity formulas used as the target of the protocol.","marker":"[29]"},{"why":"Defines the 1→3 optimal asymmetric universal cloning isometry, its parameter constraint, and the optimal fidelities used by the 1→3 protocol.","marker":"[30]"},{"why":"Supplies the classical linear multicast network coding theory and the max-flow/min-cut condition that underpins the quantum protocol.","marker":"[1, 2]"},{"why":"Shows that quantum teleportation with free classical communication effectively reverses channel direction, justifying the undirected-edge network model.","marker":"[9]"},{"why":"Establishes the no-cloning theorem that motivates formulating quantum multicast as optimal cloning.","marker":"[23]"},{"why":"Supplies the symmetric optimal clone multicast protocols that this paper extends to the asymmetric case.","marker":"[11–13]"}],"fun_headline_variants":["Asymmetric optimal clone multicast over quantum networks","Network coding for asymmetric optimal quantum clones","Constant-entanglement multicast of asymmetric optimal clones","Distributed asymmetric optimal cloning over networks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 1-to-2 protocol assumes that the partial swap $\\Gamma^{(r)}$ of Eq. (23) moves the $d$-level states $|r-1\\rangle$ and $|r\\rangle$ out of systems $E$ and $F$ into the two-dimensional systems $G$ and $H$ after Step 5, even though the swap as written acts only on the $|0\\rangle,|1\\rangle$ subspace.","fun_headline_variants_meta":{"raw":{"variants":["Asymmetric optimal clone multicast over quantum networks","Network coding for asymmetric optimal quantum clones","Constant-entanglement multicast of asymmetric optimal clones","Distributed asymmetric optimal cloning over networks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000671,"raw_usage":{"total_tokens":3027,"prompt_tokens":886,"completion_tokens":2141,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":502,"completion_tokens_details":{"reasoning_tokens":2088}},"tokens_in":502,"tokens_out":2141,"duration_ms":16638,"temperature":1.0,"reasoning_tokens":2088,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:37:29.618628+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Apply the protocol with $q=3$ and measurement outcome $r=3$, and track the component $\\cos\\eta\\,|2\\rangle_E|3\\rangle_F$ through Step 7: since $\\Gamma^{(3)}$ is the identity whenever the control register is not in $|3\\rangle$, this component stays in $EF$ and never moves to $GH$, so Eq. (34) fails, and the final reduced state on $EF$ would show a fidelity below the predicted optimal asymmetric bounds.","supporting_citations":[{"cited_title":"Per- fect quantum network communication protocol based on clas- sical network coding,","cited_arxiv_id":null,"evidence_quote":"Provides the quantum multicast subroutine used to distribute the compressed state as a GHZ-type state to all targets."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the 1→2 optimal asymmetric universal cloning isometry and the optimal fidelity formulas used as the target of the protocol."},{"cited_title":"Iblisdir, A","cited_arxiv_id":null,"evidence_quote":"Defines the 1→3 optimal asymmetric universal cloning isometry, its parameter constraint, and the optimal fidelities used by the 1→3 protocol."},{"cited_title":"Quantum Network Communication; The Butterfly and Beyond,","cited_arxiv_id":null,"evidence_quote":"Shows that quantum teleportation with free classical communication effectively reverses channel direction, justifying the undirected-edge network model."},{"cited_title":"Wootters, Wojciech H","cited_arxiv_id":null,"evidence_quote":"Establishes the no-cloning theorem that motivates formulating quantum multicast as optimal cloning."}],"review_version":1}