{"id":"7509f5f7-b013-4572-9d77-234c7412c422","arxiv_id":"2601.04998","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":6,"one_line_summary":"A collision-based quantum circuit with non-orthogonal reservoir qubits realizes complex-balanced thermalization, enabling controllable out-of-equilibrium states, dichromatic photon correlations, and finite-temperature synchronization protected by a Liouvillian exceptional point.","lead":"This paper proposes a quantum-circuit setup where a system repeatedly collides with specially engineered 'reservoir qubits' whose states are not orthogonal, producing controllable non-equilibrium thermal states. It could give quantum simulators a practical way to prepare out-of-equilibrium states at chosen temperatures, with demonstrations in photon emission and spin synchronization.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"At the flagship parameters (g=2, \\bar t=0.05) the collision map is not shown to be trace-preserving, so the reported LEP-protected QS and photon correlations may be normalization artifacts; predictive control is therefore not yet established.","rationale":"The reader's weakest assumption identifies exactly the load-bearing issue: trace preservation and CPTP validity of the QME at the application parameters. This is not a manufactured concern; the authors explicitly acknowledge both that Eq. (1) may not preserve trace and that Eq. (2) does not necessarily generate a CPTP semigroup. The paper's own supplemental data show that decreasing \\bar t does not improve QME/collision-map agreement in the QS regime, and that only weak coupling restores agreement. Since the flagship applications use g=2, the predictive-control claim is conditional on a trace-preservation check that is nowhere reported. The paper does have useful independent content: the collision model is a concrete algorithmic prescription, and Fig. 3 shows good QME/collision agreement at tbar=0.05, g=1 for the dichromatic emission setup. But the more central synchronization application lacks this verification at its stated parameters. The reader's CONDITIONAL verdict is therefore appropriate; the concern does not force a stronger verdict until a check is performed, but it does block unconditional acceptance.","tokens_in":16166,"tokens_out":4769,"duration_ms":56697,"concrete_test":"Run the exact collision map, Eq. (1), for the Fig. 4(d) parameters: g=2, \\bar t=0.05, θq=0.55, ϕ0=π/3, J=0.2, hz=2hx=1, β=1, Nq=6, and initial state |↑↓><↑↓|. Compute Tr ρ_s^{(n)} for n up to 10,000 collision steps, along with <s_x^1> and C12 with and without renormalization. If max|1 - Tr ρ_s^{(n)}| exceeds ~1% and the renormalized vs raw curves differ visibly, the LEP-protected QS is at least partly a normalization artifact. Repeat for g=0.1 with the same \\bar t; if the trace is preserved there and the curves agree, the protocol is valid only in the weak-coupling regime, not at the advertised parameters.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the circuit platform enforces trace-preserving Markovian complex-balanced thermalization. However, the collision map in Eq. (1) uses a non-unitary evolution U = e^{-iH\\bar t} because the reservoir qubit Hamiltonian H_q is non-Hermitian; the paper itself states that this map may fail to preserve trace. It then says the paper focuses on trace-preserving regimes, but no trace-preservation check is reported for the application parameters. In particular, Fig. 4 uses g=2 and \\bar t=0.05. The Supplemental Material (Fig. S2) shows that in the QS regime the agreement between collision map and QME does not improve when \\bar t is reduced, and that the dynamics deviate from the standard Born-Markov approximation; only reducing g restores agreement. Thus the QME used to identify the Liouvillian exceptional point may not describe the actual circuit at g=2. Since observables such as <s_x^1> and C12 are computed from a normalized expectation value, a drifting trace Tr ρ_s would directly change the reported correlations and synchronization. The paper does not provide the required evidence that Tr ρ_s is conserved to, say, <1% over the simulated time window, so the two applications rest on an unverified physical normalization.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a quantum-circuit platform in which a system repeatedly collides with engineered non-Hermitian reservoir qubits, described by the collision map in Eq. (1). In the short-time and weak-coupling limits the authors derive the quantum master equation (QME) of Eq. (2) with dual spectral functions γ and γ̄, and they argue that non-orthogonality of the reservoir qubit eigenstates breaks detailed balance and produces complex-balanced thermalization with a modified KMS relation. Two applications are presented: temporally correlated dichromatic photon emission from a three-level system, and Liouvillian-exceptional-point-protected quantum synchronization of two spins at finite temperature. The paper claims predictive control over out-of-equilibrium state preparation.","tokens_in":16599,"tokens_out":3550,"duration_ms":41646,"significance":"If the central claim is valid, the platform is a valuable modular construction: it connects non-Hermitian reservoir engineering, collision models, and complex-balanced thermalization, and the two applications (dichromatic photon correlations and LEP-protected synchronization) are nontrivial and potentially useful. The analytical derivation of the QME is standard in structure and the paper provides numerical comparisons between collision model and QME in the Supplemental Material, which is a strength. However, the validity of the derivation at the parameters used in the applications is not established, and the paper itself concedes that the collision map may not be trace-preserving and the QME may not generate a CPTP semigroup. Since the advertised predictive control rests on the QME/collision-map description, this is a load-bearing gap.","major_comments":[{"comment":"The collision map in Eq. (1) is not shown to be trace-preserving in the application regimes. The paper states in the main text that \"this collision map may not preserve the trace of ρ_s\" and later that the QME \"does not necessarily generate a completely positive, trace-preserving quantum dynamical semigroup.\" Yet the application figures use g=1 (Fig. 3) and g=2 (Fig. 4) with t̄=0.05, and no trace-conservation check is reported. Observables such as ⟨s_x^1⟩, C12 in Eq. (9), and G^(2) in Eq. (8) are normalized by tr_s ρ_s; if the trace drifts, these quantities are not physical expectation values. The SM (Fig. S2) explicitly shows that in the QS regime reducing t̄ does not improve collision-model/QME agreement, and only reducing g restores agreement. Therefore the QME used for the LEP analysis may not describe the actual circuit at g=2. The authors should provide quantitative data on Tr ρ_s","section":"Eq. (1), End Matter A, Fig. 4, SM Fig. S2"},{"comment":"The QME derivation assumes both t̄ ≪ 1 and g ≪ 1, but the applications use g=1 and g=2. The paper's own SM (Fig. S2) shows that for g=1 and t̄=0.05 the collision map and QME agree for the dichromatic emission, but for the synchronization setup with g=2 the agreement only improves when g is reduced, not when t̄ is reduced. This means the Liouvillian exceptional point and the associated protected quantum synchronization in Fig. 4 may be properties of the approximate QME rather than of the actual collision circuit. The authors should demonstrate the LEP (rank-2 zero-eigenvalue coalescence) directly in the collision map, or present the synchronization results in a parameter regime where the weak-coupling assumption is satisfied.","section":"End Matter A, Eq. (A2)–(A3)"},{"comment":"The derivation removes the linear-in-H_sq term by the \"stability condition\" tr_q[B^(n) ρ_q^(n)] = 0 and claims this can always be enforced by redefining B = B' − μ_b. Redefining B changes H_sq from g A⊗B' to g A⊗(B' − μ_b I) = g A⊗B' − g μ_b A⊗I, which adds an effective system Hamiltonian term. This term is not included in the QME of Eq. (2), and the claim that it \"only influences the rate towards long-term states\" is not justified. A Hamiltonian shift can affect oscillation frequencies and therefore the synchronization and LEP predictions. The authors should either include this term explicitly or prove that it does not affect the reported observables.","section":"End Matter A, stability condition"}],"minor_comments":[{"comment":"Typo: \"circ uits\" should be \"circuits\".","section":"Title"},{"comment":"The label \"Reservior\" should be \"Reservoir\".","section":"Fig. 1"},{"comment":"The notation with a dagger outside the anticommutator, { ... }†, is nonstandard and should be clarified. In particular, it is not immediately clear whether the adjoint applies to the whole anticommutator or to each term.","section":"Eq. (A2)"},{"comment":"The operator ordering in the definition of G^(2) is confusing. Please specify the normal ordering and the time arguments of p1 and p2 more precisely, especially for the cross-correlation G^(2)_12.","section":"Eq. (8)"},{"comment":"The labels \"2ndLEP\" and \"4thLEP\" are not explained in the caption or text. What defines the order of the LEP, and which one is shown in panel (a)?","section":"Fig. 4"},{"comment":"The Supplemental Material reference uses a placeholder \"URL will be inserted by publisher\". This should be completed before publication.","section":"Reference [58]"}],"recommendation":"major_revision","confidential_remarks":"The paper has an interesting core idea, but the trace-preservation problem is central and is acknowledged by the authors themselves. The SM data in Fig. S2 actually strengthen the reviewer's concern: in the synchronization regime the QME is not validated by the small-t̄ limit. I would ask the authors for explicit trace-conservation data and for either a direct verification of the LEP in the collision model or a restriction of the claims to weak coupling. If these cannot be provided, the paper's central claim of predictive control would not be supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper proposes a collision-model platform for complex-balanced thermalization using non-Hermitian reservoir qubits. The combination—biorthonormal qubit eigenstates producing dual spectral functions and a modified KMS relation—is genuinely new as far as I can tell from the citations, and the two applications are concrete and numerically worked out. The authors are also honest: they state outright that the collision map may not preserve trace and that the master equation does not necessarily generate a CPTP semigroup.\n\nThe soft spot is the mismatch between the derivation and the application parameters. The master equation is derived under g ≪ 1 and \\bar t ≪ 1, but the synchronization application runs at g = 2, and Fig. S2 shows that shrinking \\bar t at fixed g does not improve agreement between collision map and master equation in the QS regime; only reducing g does. That means the Liouvillian exceptional point, which is the basis of the LEP-protection claim, is identified from an equation that may not describe the actual circuit at those parameters. The paper even mentions an additional von-Hove limit g → 0 is needed to align the two, yet the numerics use g = 2.\n\nThe related issue is trace preservation. Observables are normalized by Tr ρ_s, so a drifting trace changes the reported correlations and synchronization amplitude. No check of Tr ρ_s versus time is reported. Since the paper says it \"focuses on\" trace-preserving regimes, this is a testable condition, but it is not demonstrated.\n\nWhat holds up: the QME derivation is standard and clearly written, and the dichromatic emission application at g = 1 shows good collision-map/QME agreement. The mechanism itself is worth pursuing. The flaws are not necessarily fatal, but the central claim of \"predictive control\" is not yet established.\n\nI would send this to peer review, but I would ask the referees to demand a systematic trace-preservation check, convergence tests at the actual parameters, and a revised claim if the QME does not reproduce the collision map in the regimes shown. As it stands, I would not cite it for the LEP result, though I'd keep an eye on a revised version.","headline":"Novel collision-model mechanism for complex-balanced thermalization, but the flagship applications run at parameters where the derived master equation is not controlled and trace preservation is unverified.","tokens_in":17011,"tokens_out":3410,"would_cite":false,"duration_ms":35432,"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":"The paper proposes a quantum-circuit protocol in which repeated collisions with engineered, non-Hermitian reservoir qubits make complex-balanced thermalization Markovian and controllable, giving access to out-of-equilibrium states at prescr","keywords":["complex-balanced thermalization","quantum circuits","reservoir engineering","non-Hermitian qubits","Kubo-Martin-Schwinger relation","Liouvillian exceptional point","quantum synchronization","dichromatic photon emission"],"falsifier":"Run the collision map at the synchronization parameters g=2, ̄t=0.05, θq=0.55 and monitor Tr ρs at every step. If the trace deviates substantially from 1 at any point, the circuit is not a trace-preserving thermalization channel, and the QME-based predictions for the photon correlations and synchronized steady state would not describe a physically realizable probability-conserving process.","tokens_in":16095,"feed_emoji":"⛛️","tokens_out":6666,"duration_ms":71620,"temperature":0.7,"pith_summary":"The paper is trying to show that a quantum circuit built from a target system repeatedly colliding with engineered reservoir qubits can realize complex-balanced thermalization in a strictly Markovian way. By using reservoir qubits whose eigenstates are not orthogonal, the circuit generates two distinct spectral functions, and their difference produces simultaneous dissipation and amplification. A modified Kubo-Martin-Schwinger relation then sets an effective temperature, so the system can be driven towards out-of-equilibrium steady states rather than ordinary Boltzmann equilibrium. The authors demonstrate the protocol on two concrete tasks: temporally correlated dichromatic photon emission and finite-temperature quantum synchronization protected by a Liouvillian exceptional point. A sympathetic reader would care because this is a route to programmable out-of-equilibrium state preparation with a clear microscopic accounting.","feed_headline":"Program out-of-equilibrium states with colliding qubits","feed_subtitle":"A quantum-circuit platform uses non-orthogonal reservoir-qubit states to drive non-uniform heating at a chosen effective temperature.","key_machinery":"Reservoir qubits with non-Hermitian Hamiltonians of the form H_q = ω(σ_x cosh θ + i σ_y sinh θ)/2. Their right eigenstates are non-orthogonal, so the transition amplitudes В_ab and В_ba are not complex conjugates. This makes the dual spectral functions γ_ω and ̄γ_ω different, producing effective dissipation and amplification rates δ_ω = ̄γ_ω − γ_ω. The modified KMS relation ̄γ_{-ω}/γ_ω = e^{-β̄ω} ties the ratio of these rates to an effective inverse temperature, turning a simple circuit collision step into a controllable thermalization engine.","core_discovery":"In the paper's own terms, the central discovery is that non-orthogonality of reservoir-qubit eigenstates is a resource: it breaks the equality between the two spectral functions γ and ̄γ that a conventional thermal reservoir would enforce, and this breaking is exactly what converts ordinary damping into balanced amplification–dissipation dynamics. Combined with Boltzmann-like preparation of the reservoir qubits, this yields a modified KMS relation that fixes an effective temperature β̄ ≠ β for the non-equilibrium steady state. The resulting quantum master equation is Markovian and, within the weak-coupling and short-collision limits, provides a solvable rate-equation description of the platf","pith_inferences":["The same non-orthogonality mechanism should transfer to any platform where non-Hermitian qubit Hamiltonians and partial traces are available, so the protocol is portable in principle beyond the specific circuit implementation shown.","The paper’s explicit caveat about non-trace-preserving collision maps suggests a testable boundary: at stronger coupling the circuit’s predictions should deviate from a true CPTP quantum channel, and locating that boundary would sharpen the range of validity.","Because the modified KMS relation derives the effective temperature from the ratio of dual spectral functions, the emitted photon correlations themselves could serve as an experimental thermometer for the effective reservoir temperature.","The LEP-protected synchronization result implies that exceptional points in Liouvillians can be used as a control knob for robust quantum synchronization, an idea that could be exported to oscillator networks and quantum metrology."],"forward_implications":["A small set of reservoir qubits can act as a programmable thermal reservoir, with the effective temperature set by the reservoir angle θq and coupling parameters rather than by the physical temperature of the environment.","The same platform can produce photon pairs with non-thermal temporal correlations—enhanced bunching at short delays and suppressed emission at long delays—suggesting a direct route to correlation-based sensing such as ghost imaging or two-photon lidar with engineered light.","Finite-temperature quantum synchronization can be made robust by sitting at a Liouvillian exceptional point, where two oscillation modes coalesce and protect the synchronized in-phase or anti-phase spin motion.","The collision model and the quantum master equation agree quantitatively in the short-collision and weak-coupling limits; the paper also notes that full agreement in some synchronization regimes requires an additional van Hove limit g→0."],"fun_headline_variants":["Non-orthogonal qubits program thermal states in quantum circuits","Colliding qubits set effective temperature via modified KMS","Balanced amplification-dissipation from qubit collisions","Quantum circuit achieves complex-balanced thermalization","Non-orthogonality enables controllable out-of-equilibrium states"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The derivation assumes weak coupling (g≪1) and a short collision interval (̄t≪1), but the applications are run with g=2 and ̄t=0.05; the paper itself notes in End Matter A that the collision map can fail to be trace-preserving and the generator may not be CPTP, so the whole predictive framework rests on the quantum master equation remaining valid in that parameter regime.","fun_headline_variants_meta":{"raw":{"variants":["Non-orthogonal qubits program thermal states in quantum circuits","Colliding qubits set effective temperature via modified KMS","Balanced amplification-dissipation from qubit collisions","Quantum circuit achieves complex-balanced thermalization","Non-orthogonality enables controllable out-of-equilibrium states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000266,"raw_usage":{"total_tokens":1400,"prompt_tokens":649,"completion_tokens":751,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":393,"completion_tokens_details":{"reasoning_tokens":673}},"tokens_in":393,"tokens_out":751,"duration_ms":8467,"temperature":1.0,"reasoning_tokens":673,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T11:48:35.534141+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the collision map at the synchronization parameters g=2, ̄t=0.05, θq=0.55 and monitor Tr ρs at every step. If the trace deviates substantially from 1 at any point, the circuit is not a trace-preserving thermalization channel, and the QME-based predictions for the photon correlations and synchronized steady state would not describe a physically realizable probability-conserving process.","supporting_citations":[],"review_version":1}