{"id":"ec02958f-9f29-4740-8a39-ab9919da4c68","arxiv_id":"2608.07695","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Open quantum walks on a ring and on two nodes prepare ensembles of Dicke and GHZ states, with a closed-form spectral gap for the Dicke walk.","lead":"This paper constructs open quantum walk protocols that prepare GHZ, W, and Dicke states using nonunitary Kraus operators, with Dicke states recovered by postselecting the walker position on a ring graph. It also derives a closed-form approximation for the convergence gap of the Dicke walk and embeds quantum trajectories into the same graph language.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unsharp-GHZ spectral gap (Eq. 19) is wrong: for eta > 1/sqrt(2) the +-sector eigenvalue eta dominates, so the true convergence rate is 1-eta, not 1-sqrt(1-eta^2).","rationale":"The reader's weakest_assumption was physical realizability of nonlocal Kraus operators and the GHZ invariant manifold. My stress-test found a different, more concrete problem: a mathematical error in the spectral analysis of the unsharp-measurement GHZ protocol. The invariant + population subspace yields an eigenvalue eta that the paper's claimed spectrum omits; for eta > 1/sqrt(2), this eigenvalue is larger than 2cs, so the stated gap and convergence time are wrong. This is not a matter of external feasibility or consensus; it is an internal inconsistency in a central quantitative claim about convergence. The ideal two-step GHZ construction and the Dicke-state Markov-chain analysis appear sound, so the paper is not fundamentally unsound, but the unsharp convergence analysis requires major revision. The concrete test, exact application of Lambda_eta to a simple operator or a short simulation, would settle the issue unambiguously. Because the error is in a claimed contribution of the abstract ('analyze how unsharp measurements affect ... the convergence of the walk'), the paper cannot be accepted without correction; hence the verdict remains CONDITIONAL, but for a different and more serious reason than the reader's.","tokens_in":15817,"tokens_out":52659,"duration_ms":466911,"concrete_test":"For eta = 0.9, apply Lambda_eta to the Hermitian operator rho_diff = |GHZ+><GHZ+| otimes (|0><0| - |1><1|) using the Kraus operators of Eq (11) with M+-. Direct calculation gives Lambda_eta(rho_diff) = 0.9 rho_diff, confirming the missing eigenvalue. Then simulate the walk from |0^N>|0> and monitor the total-variation distance to the stationary state (15); the decay envelope should follow 0.9^t, not (2cs)^t = 0.436^t, so the time to reach 1e-3 is about 690 steps, not about 8 steps.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In Sec III C the map Lambda_eta is claimed to have spectrum {1, 2cs, 0}, giving gap 1-2cs. This omits the invariant +-sector population subspace. For |GHZ+>, the transition operators M+, M-, Z1 M-, and Z1 M+ act on |+><+| as a two-node chain with transition probabilities c^2 and s^2. The 2x2 population matrix for the + sector has eigenvalues 1 and c^2 - s^2 = eta, with eigenvector |+><+| otimes (|0><0| - |1><1|). Thus eta is an eigenvalue of Lambda_eta. For eta > 1/sqrt(2), eta > 2cs, so the slowest decay rate is eta and the gap is 1-eta, not 1-2cs. Eq (19) is therefore only valid for eta < 1/sqrt(2). In the high-sharpness regime eta approx 1, the paper predicts near-instant convergence (gap near 1), whereas the + population mixing is controlled by s^2 = (1-eta)/2, giving gap approx 1-eta and a much slower relaxation. This invalidates the convergence-time analysis in Sec III C and any optimization over eta based on it.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes state-preparation protocols for entangled states based on open quantum walks with nonunitary Kraus operators. Section III gives a two-node OQW that prepares GHZ-type states: with ideal projectors (I±X^⊗N)/2 and a Z_1 edge correction, the target |GHZ+> is reached deterministically in two steps, with uniqueness only on the invariant manifold M; replacing the projectors by unsharp measurements yields a stationary ensemble whose unconditional fidelity is (1+η)/2, improves to F0>F under walker postselection, and relaxes with spectral gap 1−√(1−η^2). Section IV constructs a ring OQW with collective jump operator Jc and no-jump operator Kc that prepares an ensemble of Dicke states |D_N^{N−k}> at node k; the underlying Markov chain has a p,A-independent steady state, mean cycle time n1=2A^2H_N/[p(N+1)], and a closed-form approximate spectral gap whose large-N form gives tconv ≈ ln(C/ε)·12A^2H_N^3/[π^4p(N+1)]. Section V embeds the quantum trajectories method as a complete-graph OQW. Appendices A and B supply the Dicke relations and the Markov-chain steady-state calculation, and a public code repository is cited.","tokens_in":16064,"tokens_out":43063,"duration_ms":334173,"significance":"If correct, the paper offers a clean graph-theoretic reformulation of dissipative Dicke and GHZ state preparation, with an analytic handle on convergence time. I found the central derivations sound: the Dicke action in Eqs. (A1)–(A3), the steady state in Eq. (35), the cycle time in Eq. (40), and the spectral-gap expansion in Eqs. (48)–(51) all check out, as does the reported agreement with exact numerical diagonalization. The GHZ two-step convergence and the unsharp stationary state in Eqs. (15)–(16) are also consistent. I specifically verified the spectral claim in Sec. III C: on the invariant block-diagonal sector reachable from the walker-product initial state, the eigenvalues of Λ_η are {1, 2cs, 0}, so Eq. (19) is not afflicted by the suspected η eigenvalue; that concern does not land. The authors are also explicit about operator nonlocality and about the probabilistic Θ(N log N) overhead for balanced Dicke states, and they provide a public code repository. The protocols are more structural than immediately practical, but the paper's claims are appropriately scoped.","major_comments":[],"minor_comments":[{"comment":"The displayed expression for S2 contains a typographical error: the correct partial-fraction result is S2 = (A^4/p^2)[2H_N^{(2)} + 4H_N/(N+1)]/(N+1)^2. The final gap formula in Eq. (51) is consistent with the corrected expression, so this is a local typo, but it should be fixed.","section":"Sec. IV C, Eq. (50)"},{"comment":"The phrase 'any individual Dicke state' is too broad. The protocol generates |D_N^k> for k=1,...,N (nodes 0,...,N-1 after reset), but not the vacuum state |D_N^0>; please state this explicitly, for example as 'all Dicke states except |D_N^0>'.","section":"Abstract and Sec. IV B"},{"comment":"The sentence 'Equation (52) reproduces the observed non-monotonicity ... broad maximum near N≃17' is misleading as printed. Equation (52) is the large-N surrogate H_N^3/(N+1), which peaks near N≃12, as the next sentence itself notes; the maximum near 17 comes from the fuller Eq. (51). The attribution should be corrected.","section":"Sec. IV C, text after Eq. (52)"},{"comment":"The notation 'v(k)=p k' should read 'v(k)=p_k' to match the hop probability defined immediately before.","section":"Sec. IV B, Eq. (36)"},{"comment":"The fitted equations display '¡' in place of a minus sign; please correct the typography in the captions.","section":"Figs. 6 and 7 captions"}],"recommendation":"minor_revision","confidential_remarks":"To the editor: this is a sound theoretical contribution within the journal's scope. The main limitation is the nonlocality of the engineered Kraus operators, which the authors disclose; whether that restricts practical impact is an editorial judgment. I saw no circularity and no unsupported novelty claims. The minor issues listed above should be straightforward to address."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid, honest paper, and the one thing I was told to worry about—the unsharp GHZ spectral gap—doesn't survive contact with the actual transition operators. The Z1 correction on the edge |1>→|0> sends |+> to |->, so the '+' sector is not a closed two-node chain. I recomputed the population transition matrix in the GHZ basis: it has eigenvalues 1 and 0, with η appearing in the stationary eigenvector, not as a separate eigenvalue. The coherence sector is what supplies the 2cs eigenvalue, consistent with Eq. (19). So the convergence analysis in Sec. III C stands, at least on this point.\n\nWhat is genuinely new: the two-node GHZ walk with stabilizer Kraus operators, the ring-shaped Dicke walk with collective jump operators, and the closed-form approximate spectral gap for the Dicke ring. I checked the steady state and cycle time in Appendix B and the gap expansion against the transition matrix; they are correct. The paper is also unusually transparent. It states plainly that the Kraus operators are nonlocal, that the Dicke scheme is probabilistic with Θ(N log N) repetitions for balanced states, and that the quantum-trajectories embedding is a reformulation rather than a speedup. That honesty is real and worth crediting.\n\nSoft spots, in proportion: the abstract's \"any individual Dicke state\" is too strong. The protocol produces D_N through D_1, but never D_0, because the walker lives on N nodes and the reset edge maps D_1 back to D_N. That is an easy fix, but it should be made. The code repository should be pinned to a commit hash; minor. The larger practical ceiling is the nonlocality of the engineered dissipation, and the paper concedes this directly.\n\nVerdict: for anyone working on dissipative state preparation or open quantum walks, this deserves serious referee time. I would accept it after minor revision. The central claims are supported; the flaws are in scope claims and presentation, not in the math.","headline":"Worth a serious referee: the GHZ and Dicke OQW constructions are new, the appendix derivations check out, and the flagged spectral-gap concern evaporates once you include the Z1 edge correction.","tokens_in":16586,"tokens_out":12097,"would_cite":true,"duration_ms":104168,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Open quantum walks with engineered nonunitary Kraus operators prepare W, Dicke, and GHZ states, with a closed-form spectral gap governing Dicke convergence.","keywords":["open quantum walks","state preparation","Dicke states","W states","GHZ states","dissipative quantum computation","spectral gap","quantum trajectories"],"falsifier":"Run the corrected two-node walk on $N$ qubits from $|0^N\\rangle|0\\rangle$ and measure the graph and internal register after two steps: the claim predicts the state $|GHZ_+\\rangle\\langle GHZ_+|\\otimes|0\\rangle\\langle 0|$ with probability one. Observing any weight on node 1, or any internal component outside the $\\pm1$ eigenspaces of $X^{\\otimes N}$ (beyond numerical error), would falsify deterministic GHZ preparation.","tokens_in":15606,"feed_emoji":"⚛️","tokens_out":13464,"duration_ms":103019,"temperature":0.7,"pith_summary":"This paper claims that open quantum walks—quantum walks whose motion is driven entirely by engineered dissipation—can serve as a state-preparation framework for entangled states when the Kraus operators are allowed to be nonunitary and graph-dependent. The authors exhibit a two-node walk whose edge operators are the projectors $K_{\\pm}=(I\\pm X^{\\otimes N})/2$ of the global stabilizer $X^{\\otimes N}$: with a $Z_1$ correction on one edge it prepares $|GHZ_+\\rangle$ deterministically in two steps, with the graph position providing the outcome record. They also exhibit a ring-shaped walk driven by the collective jump $J_c=(J_1+\\cdots+J_N)/A$ and its no-jump partner $K_c=\\sqrt{I-J_c^\\dagger J_c}$, whose steady state is an ensemble of Dicke states with weights independent of $p$ and $A$; postselecting the walker at node $k$ yields $|D_N^{N-k}\\rangle$, so the W state is recovered at node $N-1$. For this walk they derive a closed-form approximate spectral gap that reproduces the numerically observed convergence time, including its non-monotonic dependence on $N$. The paper further shows that the quantum trajectories method is exactly an open quantum walk with vertex-independent Kraus operators on a complete graph, embedding it as a collision model. The protocols' economy in walk steps is real but is purchased with explicitly nonlocal, $N$-body Kraus operators, a cost the paper states plainly.","feed_headline":"Dissipative graph walks prepare W, Dicke, GHZ states","feed_subtitle":"The walker's position labels the entangled state, and a closed-form gap sets the convergence time.","key_machinery":"The central object is the open quantum walk itself, specified by edge operators $M^i_j=B^i_j\\otimes|j\\rangle\\langle i|$ with the completeness condition $\\sum_j B^{j\\dagger}_i B^j_i=I$; the graph degree of freedom simultaneously drives the dissipation and acts as a classical record of the outcome. For the GHZ protocol the load-bearing operators are the stabilizer projectors $K_\\pm=(I\\pm X^{\\otimes N})/2$, with the unitary correction $Z_1$ inserted on the $|1\\rangle\\to|0\\rangle$ edge to break the periodicity of the unconditioned walk, and the invariant manifold $\\mathcal{M}=\\mathrm{span}\\{|0^N\\rangle,|1^N\\rangle\\}$ that contains the initial state. For the Dicke protocol the load-bearing pair is the collective jump $J_c=(J_1+\\cdots+J_N)/A$ and the no-jump operator $K_c=\\sqrt{I-J_c^\\dagger J_c}$, whose actions on Dicke states are $J_c|D_N^k\\rangle=\\sqrt{kp(N-k+1)/A^2}|D_N^{k-1}\\rangle$ and $K_c|D_N^k\\rangle=\\sqrt{1-kp(N-k+1)/A^2}|D_N^k\\rangle$; these turn the ring graph into a ladder in excitation number. The convergence analysis rests on the induced cyclic bidiagonal Markov chain with hop probability $p_k=p(N-k)(k+1)/A^2$, whose characteristic equation yields the approximate spectral gap and hence the relaxation time.","core_discovery":"On its own terms, the paper's central discovery is that open quantum walks with nonunitary, graph-dependent Kraus operators constitute a state-preparation language broad enough to cover the three prototypical entangled resources of quantum communication and distributed computation. For GHZ states, the pair $K_{\\pm}=(I\\pm X^{\\otimes N})/2$ defines a two-node walk whose first step splits the register into $|GHZ_+\\rangle$ at node 0 and $|GHZ_-\\rangle$ at node 1; inserting the correction $Z_1$ into the edge $|1\\rangle\\to|0\\rangle$ makes the walk reach $|GHZ_+\\rangle\\langle GHZ_+|\\otimes|0\\rangle\\langle 0|$ in exactly two steps and stay there, with uniqueness of this attractor holding inside the invariant manifold $\\mathcal{M}=\\mathrm{span}\\{|0^N\\rangle,|1^N\\rangle\\}$. For Dicke states, the ring walk with jump $J_c$ and no-jump $K_c$ sends the node label to the excitation sector: after $m$ steps the internal state is $\\sum_k p^{(m)}_k |D_N^{N-k}\\rangle\\langle D_N^{N-k}|\\otimes|k\\rangle\\langle k|$, and measuring the graph gives the corresponding Dicke state; the steady-state node distribution is $\\pi_k=2H_N/[(N+1)(N-k)(k+1)]$, independent of $p$ and $A$, and the subdominant eigenvalues of the transition matrix yield the approximate gap $\\Delta\\simeq \\pi^2 p[2(N+1)H_N^{(2)}+4H_N]/(4A^2H_N^3)$. The paper also establishes that the quantum trajectories method is the OQW with $M^i_j=K_j\\otimes|j\\rangle\\langle i|$ and vertex-independent $K_j$, which is structurally a collision model.","pith_inferences":["Inference: the parameter-independence of the Dicke steady state suggests a robustness test—if the collective decay rates drift during a run, the long-time node distribution should remain $\\pi_k$, which would make the ensemble-preparation protocol self-calibrating in a way unitary circuits are not.","Inference: the closed-form gap for a cyclic bidiagonal chain is a transferable result; similar slowly varying birth-death chains appear in classical load-balancing and population dynamics, where the same second-order expansion could estimate mixing times.","Inference: because the GHZ claim is restricted to $\\mathcal{M}$, a natural extension the paper does not pursue is to add a second stabilizer measurement that projects the full Hilbert space onto a unique attractor, at the price of an additional graph register.","Inference: the nonlocality caveat implies that a practical implementation would likely need a shared bosonic mode for $J_c$ and an ancilla-assisted readout for $X^{\\otimes N}$; the paper names these resources but does not quantify their overhead."],"forward_implications":["The GHZ protocol produces $|GHZ_+\\rangle$ deterministically in two walk steps for any $N$, without measuring the walker, provided the evolution is confined to the invariant manifold $\\mathcal{M}$.","The Dicke protocol prepares any desired Dicke state $|D_N^{N-k}\\rangle$ by postselecting node $k$, with success probability $\\pi_k$; the W state at node $N-1$ needs about $2H_N$ repetitions, a logarithmic overhead.","The steady-state node distribution of the Dicke walk is independent of the decay rates $p$ and $A$, so the long-time ensemble is fixed once the graph size is fixed, even though the transient dynamics depends on both parameters.","With unsharp measurements of sharpness $\\eta$, the unconditional GHZ fidelity is $(1+\\eta)/2$ and the spectral gap is $1-\\sqrt{1-\\eta^2}$, so weak measurements slow convergence quadratically.","Since the quantum trajectories method is an OQW with vertex-independent Kraus operators, any Lindblad evolution discretized by Kraus operators admits a collision-model graph representation."],"supporting_citations":[{"why":"Defines open quantum walks and proves the block-diagonal graph structure used in every protocol.","marker":"[4]"},{"why":"Supplies the linear OQW model, the diagram and notation conventions, and the Markov-chain viewpoint used for the Dicke walk.","marker":"[6]"},{"why":"Establishes the dissipative-computation paradigm that the protocols aim to extend, and serves as the quasi-locality baseline.","marker":"[1]"},{"why":"Introduces dissipative entangled-state preparation via quantum Markov processes, the paradigm the OQW protocols are compared against.","marker":"[25]"},{"why":"Provides the deterministic Dicke-state circuits with O(N) depth that the probabilistic OQW protocol is explicitly compared with.","marker":"[16]"},{"why":"Proposes dissipative preparation of large W states in optical cavities, the comparison for the W-state case and the locality discussion.","marker":"[29]"},{"why":"Originates collision models, which the quantum-trajectories embedding is identified with.","marker":"[33]"},{"why":"Supplies the quantum trajectories method that the OQW construction reproduces.","marker":"[46]"}],"fun_headline_variants":["Open walks turn graphs into W, Dicke, GHZ factories","Walker's position picks the entangled state","Nonunitary quantum walks prepare W, Dicke, GHZ","One graph walk, three canonical entangled states","Graph walks: a route to W, Dicke, and GHZ states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the $N$-body, nonlocal Kraus operators at the heart of the protocols—$K_\\pm=(I\\pm X^{\\otimes N})/2$ and $J_c=(J_1+\\cdots+J_N)/A$ with $K_c=\\sqrt{I-J_c^\\dagger J_c}$—can be physically realized as engineered dissipation, and for the GHZ protocol that the evolution never leaves the invariant manifold $\\mathcal{M}=\\mathrm{span}\\{|0^N\\rangle,|1^N\\rangle\\}$.","fun_headline_variants_meta":{"raw":{"variants":["Open walks turn graphs into W, Dicke, GHZ factories","Walker's position picks the entangled state","Nonunitary quantum walks prepare W, Dicke, GHZ","One graph walk, three canonical entangled states","Graph walks: a route to W, Dicke, and GHZ states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001731,"raw_usage":{"total_tokens":6917,"prompt_tokens":1092,"completion_tokens":5825,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":708,"completion_tokens_details":{"reasoning_tokens":5743}},"tokens_in":708,"tokens_out":5825,"duration_ms":40521,"temperature":1.0,"reasoning_tokens":5743,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T00:27:07.990877+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the corrected two-node walk on $N$ qubits from $|0^N\\rangle|0\\rangle$ and measure the graph and internal register after two steps: the claim predicts the state $|GHZ_+\\rangle\\langle GHZ_+|\\otimes|0\\rangle\\langle 0|$ with probability one. Observing any weight on node 1, or any internal component outside the $\\pm1$ eigenspaces of $X^{\\otimes N}$ (beyond numerical error), would falsify deterministic GHZ preparation.","supporting_citations":[{"cited_title":"efficient","cited_arxiv_id":null,"evidence_quote":"Defines open quantum walks and proves the block-diagonal graph structure used in every protocol."},{"cited_title":"Korkmaz, D","cited_arxiv_id":null,"evidence_quote":"Supplies the linear OQW model, the diagram and notation conventions, and the Markov-chain viewpoint used for the Dicke walk."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the dissipative-computation paradigm that the protocols aim to extend, and serves as the quasi-locality baseline."},{"cited_title":"Hua, Y.-R","cited_arxiv_id":null,"evidence_quote":"Provides the deterministic Dicke-state circuits with O(N) depth that the probabilistic OQW protocol is explicitly compared with."},{"cited_title":"Sweke, I","cited_arxiv_id":null,"evidence_quote":"Originates collision models, which the quantum-trajectories embedding is identified with."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the quantum trajectories method that the OQW construction reproduces."}],"review_version":1}