{"id":"4c5248ca-7dc9-4ed1-828b-f6e8996a74c5","arxiv_id":"1908.04124","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Lazy open quantum walks on a d-dimensional lattice converge to a Gaussian distribution, with an explicit covariance formula that matches a direct numerical simulation.","lead":"This paper adds a stay-in-place option to open quantum walks, random walks driven by a noisy environment, and proves their long-run position distribution is a Gaussian. It matters because any open quantum walk derived from a real physical setup necessarily includes this stay-in-place option, so the result applies to physically realistic models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The CLT is asserted via [28], but the lazy case changes the transition kernel and the Poisson equation; the key conditional-variance convergence in Eq. (34) is not demonstrated, so formula (35) rests on an unverified martingale-CLT condition.","rationale":"The reader correctly identifies uniqueness of the steady state as an explicitly assumed and important hypothesis, and the authors themselves flag it as a limitation. I agree that uniqueness is necessary for the stated form of the CLT, since without it the Cesaro limit in (22) is a random steady state and the covariance in (35) is not well defined. However, the more load-bearing gap is that even on the stated domain of the theorem, the proof of the CLT is incomplete: the paper asserts, rather than verifies, the conditional-variance convergence (34) for the lazy chain. The numerical example in Section IIIE checks formula (35) for a single family of operators and shows consistency, but it does not establish the martingale condition needed for the theorem; a single numerical check cannot rule out failure of (34) in other lazy OQWs. The paper is a plausible extension of [28], and the evidence given is suggestive, but the central theorem is not yet fully supported. A CONDITIONAL verdict is therefore appropriate: accept only if the martingale CLT conditions, especially (34), are verified for the lazy case, or if a precise reference establishes that [28]'s proof covers the additional A0 term without further assumptions.","tokens_in":18269,"tokens_out":14697,"duration_ms":166243,"concrete_test":"Simulate the quantum trajectory chain of Section IIB for a generic lazy OQW with a unique, non-degenerate steady state (for example, the operators in Eq. (18) perturbed by a small random Hermitian term, or the microscopic model of Example 2). Along one long trajectory, compute the exact conditional variance at each step from the known jump distribution and form the running average b_n = (1/n) sum_{k=1}^n E[(Delta M_k)^2 | F_{k-1}] using Eq. (27). Compare b_n with sigma_l^2 obtained from formula (35) for several directions l, and repeat from several initial internal states tau_0. If b_n does not converge to sigma_l^2, or if the limit depends on tau_0, then condition (34) is not implied by (22) and uniqueness, and the central CLT claim needs a revised proof.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is the martingale CLT for homogeneous lazy OQWs with a unique steady state. The proof reduces (X_n - nm)/sqrt(n) to M_n/sqrt(n) plus a bounded term, then invokes the martingale CLT. The boundedness needed for Lindeberg condition (33) is straightforward, but the other condition, convergence of the conditional variance in (34), is the load-bearing step. The paper says 'one can show' (34) using the ergodic theorem (22) and cites [28]. However, (22) is only a.s. Cesaro convergence of the trajectory internal states tau_j to rho_infinity. The integrand in (34) is a nonlinear function of tau_{k-1}: it contains p_j = Tr(A_j tau A_j^dagger) and the normalized post-jump states A_j tau A_j^dagger / p_j. Cesaro convergence of tau_j alone does not imply convergence of Cesaro averages of such nonlinear functions unless an additional ergodicity or mixing property of the (tau_n, Delta X_n) chain is established. Moreover, including A0 changes the transition operator in (23) and the dual map L^dagger in (28a), so the operators L_l and the covariance formula (35) are not the identical objects as in [28]; the estimates proving (34) in [28] do not automatically transfer. If (34) fails, the asymptotic covariance can differ from (35), or the Gaussian limit may not exist at all. The unique-steady-state assumption is necessary but is not by itself shown to be sufficient; the paper supplies no additional condition that guarantees (34).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends the discrete-time open quantum walk by adding a 'lazy' Kraus operator A0 that leaves the lattice site unchanged. For a homogeneous lazy OQW on Z^d with a unique steady state, it claims a central limit theorem: (X_n - n m)/sqrt(n) converges in distribution to a Gaussian, with mean m and covariance matrix C_ij given by Eq. (35). The derivation follows [28]: a Doob decomposition turns the centered position into a martingale plus a bounded term, and a martingale CLT is invoked. The paper also connects the lazy map to a microscopic master equation, supplies analytic variance formulas for one- and two-dimensional microscopic models, and reports a numerical simulation (Table I) confirming the variance formula for a specific one-dimensional walk.","tokens_in":18596,"tokens_out":12028,"duration_ms":129909,"significance":"If the CLT is established, the paper fills a physically motivated gap: [40] showed that any microscopically derived OQW must contain a self-jumping term, so the lazy version is the relevant object for applications. The variance formula (35) is concrete and falsifiable, and the numerical check in Section III E is an independent test with no fitted parameters; the analytic examples in Sections III C and III D also reproduce the microscopic expressions of [40], which supports the internal consistency of the calculation. The main weakness is that the proof of the martingale condition (34) is not supplied, so the central result is currently an assertion rather than a demonstrated theorem; the numerical and analytic checks make the claim plausible but do not replace the missing verification.","major_comments":[{"comment":"Equation (34) is the load-bearing condition for the martingale CLT, but it is not proved. The integrand E[(Delta M_k)^2 | F_{k-1}] is a nonlinear function of tau_{k-1} (it involves p_j = Tr(A_j tau_{k-1} A_j^dag) and the normalized post-jump states A_j tau_{k-1} A_j^dag / p_j). The cited ergodic theorem (22) only asserts almost-sure Cesaro convergence of the internal states tau_j to rho_infty; this alone does not imply convergence of Cesaro averages of nonlinear functions of the full chain (tau_n, Delta X_n) unless an ergodic theorem for that chain is stated and its hypotheses (e.g., irreducibility/aperiodicity or a unique invariant measure for the trajectory chain) are verified. The assertion 'one can show' therefore hides the main step, and formula (35) is not rigorously derived.","section":"III A, Eqs. (33)-(34)"},{"comment":"The statement that the increments Y_k are i.i.d. once the system is in the steady state is not correct for the quantum trajectory chain. Even with the marginal law of tau_n equal to rho_infty, the increment Delta X_n has conditional distribution P(j,n) = Tr(A_j tau_n A_j^dag) that depends on the random internal state tau_n, so successive increments are dependent. The i.i.d. setup in Eqs. (24)-(26) is also inconsistent with the martingale decomposition that follows: if the increments were i.i.d., the covariance in (35) would reduce to sum_j p_j e_j e_j^T - m m^T and the L-dependent terms would be absent. The theorem should be formulated for the stationary Markov chain used in [28].","section":"III A, Eqs. (24)-(26)"},{"comment":"The Poisson equation (28a) and the covariance formula (35) are taken from [28], but the lazy walk changes both L and L^dag through the extra operator A0. Consequently the existence of L_l and the convergence of the conditional variance are not automatic consequences of the non-lazy theorem; the manuscript must either state the exact theorem from [28] in a form covering the present transition kernel and verify its hypotheses, or give a self-contained proof. The current one-sentence justification of (34) does not do either. Additionally, the uniqueness of the fixed point of L is assumed, but no condition ensuring ergodicity of the trajectory chain is supplied, so uniqueness by itself is not shown to be sufficient.","section":"III A, Eqs. (28a) and (35)"}],"minor_comments":[{"comment":"The vectorized left-hand side of Eq. (28a) should be (I - sum_{i=0}^{2d} A_i^dag tensor A_i^T) vec(L_l); the sum as written starts at i=1 and omits A0, although L^dag includes i=0 and the numerical matrix in Eq. (87) includes the A0 contribution.","section":"III A, Eq. (31)"},{"comment":"The phrase 'l2(C) is the space of square integrable functions' should read 'the space of square-summable sequences on Z^d'; also 'posseses' should be 'possesses'.","section":"II A"},{"comment":"In Eqs. (66)-(67), both B_y and C_y are defined with sigma_z, but C_y is later used in the combination r_- = gamma_+^y - gamma_-^y as a backward transition; the relation of these operators to the positive and negative y-directions e_2 and e_4 should be stated explicitly.","section":"III D, Example 2"}],"recommendation":"major_revision","confidential_remarks":"The gap in the proof of the CLT is the reason for major revision; I do not see circularity or fabricated results, and the numerical check is honest. If the authors can either import a precise theorem from [28] that covers the lazy kernel or prove (34), I would support publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Main thing: the paper extends the OQW central limit theorem to lazy walks where a self-loop A0 is allowed, and gives an explicit covariance formula (35). The physical motivation is solid: [40] showed microscopically derived OQWs must have a self-jump term. The explicit treatment of A0 in the variance is not in [28], and the numerical check in Section III.E is genuine and convincing: the simulated variance converges to the predicted value as n grows, with no fitting constants. The connection to the microscopic GKSL form is also useful.\n\nThe soft spot is the proof. The martingale CLT is not actually supplied. Condition (33) is fine, but the load-bearing condition (34) is dismissed with \"one can show\" and a reference to the ergodic theorem (22). The stress-test note is right: Cesaro convergence of tau_n to rho_infinity does not imply convergence of the nonlinear functions in the conditional variance, and the lazy transition kernel differs, so the estimates from [28] do not automatically transfer. I suspect it can be fixed, but as written the central theorem is asserted, not proved. The relation to [35], which already proved a CLT for homogeneous OQWs on Z^d, is also not resolved: if [35] covers arbitrary finite jumps including zero displacement, then the CLT is not new; only the variance formula and the microscopic connection are.\n\nUnique steady state is assumed and flagged, which is fine. The algebraic checks in the microscopic examples are consistent, though they check the variance formula rather than the CLT itself.\n\nOverall: a clearly written, physically motivated paper for the OQW community, but the main theorem is incomplete and novelty relative to [35] is unclear. It deserves a serious referee, not a desk rejection; acceptance should require filling the gap in (34) and clarifying the prior literature.","headline":"Useful extension of the OQW CLT to lazy walks, with a checkable covariance formula and good numerics, but the proof of the martingale condition is sketched and the relation to [35] is not resolved.","tokens_in":776,"tokens_out":1338,"would_cite":false,"duration_ms":82615,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60F05","60G42","81S22","82C41"],"pacs":["03.65.Yz","05.40.Fb","02.50.Ga"],"model":"deepseek-v4-flash","headline":"For a homogeneous lazy open quantum walk on a d-dimensional lattice, if the steady state is unique, the walker's rescaled position converges in distribution to a Gaussian with an explicit covariance matrix.","keywords":["lazy open quantum walk","central limit theorem","quantum trajectories","martingale","covariance matrix","dissipative quantum walk","lattice walk","steady state"],"falsifier":"Take the lazy open quantum walk of the numerical example, where the theoretical variance is $C = 1.04167$, simulate a large number of independent trajectories, and test the empirical rescaled positions $(X_n - n m)/\\sqrt{n}$ against the Gaussian $N(0,C)$ with a standard goodness-of-fit test; a discrepancy that does not shrink as $n$ grows would refute the theorem. A complementary check is to build a homogeneous lazy walk whose operator algebra decouples into two sectors, each with its own steady state, and verify that the rescaled position fails to converge to a single Gaussian, confirming that the uniqueness assumption is essential.","tokens_in":18069,"feed_emoji":"⚛️","tokens_out":8096,"duration_ms":73815,"temperature":0.7,"pith_summary":"Open quantum walks are random walks on a graph driven entirely by dissipation: each step applies a completely positive map that updates the walker's internal quantum state and moves it between sites. Microscopically realistic derivations force an extra possibility, that the walker stays on the same site, and this paper extends the standard map to include that 'lazy' operator. The paper's central claim is a central limit theorem: for a homogeneous lazy walk on the lattice $\\mathbb{Z}^d$ with a unique steady state, the walker's position after $n$ steps, minus its mean and divided by $\\sqrt{n}$, converges in distribution to a Gaussian. It also provides a closed analytic formula for the covariance matrix of that Gaussian in terms of the steady state and the jump operators. The authors check the formula on three microscopic models and against a numerical simulation, finding agreement.","feed_headline":"Lazy open quantum walks obey a central limit theorem","feed_subtitle":"Even a walker that may stay put, as microscopic dynamics requires, converges to a Gaussian with explicit covariance.","key_machinery":"The load-bearing mechanism is the martingale decomposition of the position process in the quantum-trajectory representation, together with the operator $L_l$ solving equation (28a). The difference between the walker's position and its mean is written as a martingale $M_n$ plus a predictable process whose magnitude is bounded uniformly in $n$, so the asymptotic law is carried entirely by the martingale. The martingale central limit theorem applies once its two conditions are checked: the jumps have uniformly bounded size, and the conditional variance converges to $\\sigma_l^2 = l^T C l$ via the ergodic theorem for the unique steady state. The analytic output is the covariance formula (35), and the genuinely new ingredient is the operator $A_0$ encoding the walker's self-jump, which changes the steady state, the mean, and the $L_i$ matrices.","core_discovery":"The central claim is that a homogeneous lazy open quantum walk on $\\mathbb{Z}^d$, whenever the completely positive map $L(\\tau)=\\sum_{j=0}^{2d} A_j \\tau A_j^\\dagger$ has a unique steady state $\\rho_\\infty$, satisfies $(X_n - n m)/\\sqrt{n}$ converging in distribution to a centered Gaussian with covariance $C$, where $m$ is the mean jump vector and $C$ is the explicit expression in formula (35). The proof passes to the quantum-trajectory picture, where the walk becomes a classical Markov chain, and decomposes the position process into a martingale plus a bounded predictable part. Verifying the two martingale central limit theorem conditions, with the quadratic variation controlled by the ergodic theorem for the unique steady state, yields the Gaussian law. Because the lazy operator $A_0$ enters both the map and its dual, the steady state $\\rho_\\infty$, the mean $m$, and the $L_i$ operators defined by $L_i - L^\\dagger(L_i) = \\tilde A_i - m_i I$ all differ from their non-lazy counterparts, so formula (35) genuinely describes the lazy model. A key added observation is that the system of equations for the $L_i$ is degenerate, with solutions unique up to a multiple of the identity.","pith_inferences":["If the uniqueness assumption fails but the steady states form a convex set, a plausible extension is a conditional central limit theorem: the limiting law would be a mixture of Gaussians weighted by the initial state's projection onto each ergodic sector, with the same covariance formula applied sector by sector.","The Fourier-space dual method used for non-lazy open quantum walks could be applied to the lazy map, offering an independent derivation of formula (35) and possibly stronger local limit theorems for these walks.","Because the self-jump probability can be tuned experimentally, formula (35) suggests a direct test in optical-lattice or trapped-ion implementations: vary the lazy operator and compare the measured diffusion constant with the analytic prediction."],"forward_implications":["Every open quantum walk that comes from a microscopic derivation includes a self-jump term, so the theorem applies to all such walks on a lattice: their long-time position distribution is Gaussian whenever the steady state is unique.","The covariance matrix of that Gaussian can be computed directly from $\\rho_\\infty$ and the $L_i$ operators, giving an analytic prediction for the diffusion tensor of a dissipative walker.","Ignoring the self-jump systematically changes the predicted mean and variance, so comparisons to experiment must use the lazy formulas rather than the non-lazy ones.","The degeneracy of the $L_i$ equation, with uniqueness up to a multiple of the identity, simplifies the task of solving for the covariance in real models.","In the microscopic limit, both the steady-state equation and the evolution equation for $L_l$ become independent of the time step and take a standard dissipative master-equation form, connecting the discrete-time theorem to the continuous-time master equation."],"supporting_citations":[{"why":"Supplies the original central limit theorem for non-lazy open quantum walks and the martingale method that this paper extends to include the self-jump operator.","marker":"[28]"},{"why":"Provides the microscopic derivation showing that every physically consistent open quantum walk must include a self-jumping term, motivating the lazy model and supplying the models used for analytic checks.","marker":"[40]"},{"why":"Establishes the almost-sure ergodic convergence of the trajectory-averaged state to the steady state, which the proof uses to identify the martingale's limiting variance.","marker":"[27]"},{"why":"States the martingale central limit theorem whose two conditions are verified in the paper.","marker":"[43]"},{"why":"Provides the decomposition used to split the position process into a martingale and a bounded predictable process.","marker":"[42]"},{"why":"Gives the standard Markovian master-equation formalism used to rewrite the steady-state and $L_l$ equations in the microscopic limit.","marker":"[47]"}],"fun_headline_variants":["Lazy open quantum walks still hit Gaussian limit","Central limit theorem proven for lazy open quantum walks","Even lazy quantum walks converge to Gaussian","Lazy OQWs: explicit covariance from CLT"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire result rests on the assumption that the map $L(\\tau)=\\sum_{j=0}^{2d} A_j \\tau A_j^\\dagger$ has exactly one steady state $\\rho_\\infty$; if several steady states coexist, the ergodic convergence and the covariance formula are not justified, and the authors explicitly note that the theorem may then fail.","fun_headline_variants_meta":{"raw":{"variants":["Lazy open quantum walks still hit Gaussian limit","Central limit theorem proven for lazy open quantum walks","Even lazy quantum walks converge to Gaussian","Lazy OQWs: explicit covariance from CLT"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000437,"raw_usage":{"total_tokens":2215,"prompt_tokens":929,"completion_tokens":1286,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":545,"completion_tokens_details":{"reasoning_tokens":1227}},"tokens_in":545,"tokens_out":1286,"duration_ms":10161,"temperature":1.0,"reasoning_tokens":1227,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:51:16.036622+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the lazy open quantum walk of the numerical example, where the theoretical variance is $C = 1.04167$, simulate a large number of independent trajectories, and test the empirical rescaled positions $(X_n - n m)/\\sqrt{n}$ against the Gaussian $N(0,C)$ with a standard goodness-of-fit test; a discrepancy that does not shrink as $n$ grows would refute the theorem. A complementary check is to build a homogeneous lazy walk whose operator algebra decouples into two sectors, each with its own steady state, and verify that the rescaled position fails to converge to a single Gaussian, confirming that the uniqueness assumption is essential.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the original central limit theorem for non-lazy open quantum walks and the martingale method that this paper extends to include the self-jump operator."},{"cited_title":"Sinayskiy and F","cited_arxiv_id":null,"evidence_quote":"Provides the microscopic derivation showing that every physically consistent open quantum walk must include a self-jumping term, motivating the lazy model and supplying the models used for analytic checks."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the almost-sure ergodic convergence of the trajectory-averaged state to the steady state, which the proof uses to identify the martingale's limiting variance."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"States the martingale central limit theorem whose two conditions are verified in the paper."},{"cited_title":"Carbone and Y","cited_arxiv_id":null,"evidence_quote":"Provides the decomposition used to split the position process into a martingale and a bounded predictable process."},{"cited_title":"Sinayskiy and F","cited_arxiv_id":null,"evidence_quote":"Gives the standard Markovian master-equation formalism used to rewrite the steady-state and $L_l$ equations in the microscopic limit."}],"review_version":1}