{"id":"5ad4202f-0f86-4f2d-95f1-637c40ea1e8a","arxiv_id":"2607.11055","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":5,"one_line_summary":"A linear integral equation solved by matrix-free QTT-GMRES yields vertex-corrected nonequilibrium correlation functions, revealing enhanced order-parameter fluctuations near nonthermal criticality in the quenched Hubbard model.","lead":"A new computational method evaluates vertex-corrected two-particle correlation functions far from equilibrium by combining a contour-dependent virtual probe with a matrix-free quantics-tensor-train solver. The approach shows that collective fluctuations of an order parameter grow near a nonthermal critical point and that vertex corrections qualitatively change the dynamics.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The paper's strongest claim has two parts: (i) a practical computational route to vertex-corrected nonequilibrium two-particle functions via the linear equation (3) and a matrix-free QTT-GMRES solver, and (ii) the concrete observation that those vertex corrections reverse the short-time fluctuation dynamics and that both amplitude and lifetime grow near the nonthermal critical point. Part (i) is cleanly derived, benchmarked at short times against the more expensive GCH pump-probe route, and formulated so that it applies to any self-energy for which the Jacobian-vector product Σ'_int[G,G'] can be evaluated. Part (ii) is obtained inside a well-documented second-order DMFT setup already used for the same quench; the qualitative reversal is traced to the Hartree–Fock piece of the kernel (Fig. 2(d) and the RPA analysis in the End Matter), which is the leading Stoner contribution and is expected to survive in higher-order treatments. The reader's weakest assumption correctly notes that a non-perturbative impurity solver could shift U_nc^c and the quantitative size of the corrections, but that is an expected limitation of the application, not a flaw that undermines either the method or the qualitative message the authors draw. The acknowledged t_max=32 ceiling is a practical constraint, not an inconsistency. Public code libraries further lower the barrier to independent checks. Consequently the ACCEPT verdict stands; no adjustment is warranted.","tokens_in":25237,"tokens_out":692,"duration_ms":5877,"concrete_test":"Re-run the U_i=2.0→U_f=1.2 quench with only the Hartree–Fock vertex retained (already shown in Fig. 2(d) for short times) out to t≈32 and extract τ_χ; if the growth of τ_χ and σ_m^max toward U_nc^c survives at the same qualitative level as the full second-order vertex, the demonstration is robust to the precise content of the higher diagrams within this truncation class.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption (second-order Born truncation of the impurity self-energy) is real but not load-bearing for the paper's central claim. The methodological advance—the linear integral equation (3) for G' that avoids constructing the four-time kernel K, solved matrix-free with QTTs—is independent of the particular self-energy approximation. The End Matter already shows how Σ'_int is obtained for denser kernels (e.g., GW via JVP) without storing K. The physical demonstration (vertex corrections reverse the short-time sign of σ_m and both peak amplitude and decay time grow near the nonthermal critical point) is presented as an application of that method within a controlled, previously studied second-order DMFT setup; the authors themselves flag the t_max=32 GMRES limit and the need for longer-time analysis. Short-time benchmarks against GCH pump-probe simulations (End Matter Fig. 4) and the RPA argument for the Hartree–Fock-driven short-time drop further support the qualitative message. No internal inconsistency or hidden assumption that would overturn the strongest claim was found.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript develops a generalized Keldysh framework for nonequilibrium two-particle correlation functions by introducing a contour-dependent virtual probe field. From the nonequilibrium Dyson equation it derives a linear integral equation (Eq. 3) for the Green’s-function variation G′ that incorporates vertex corrections without assembling the four-time kernel K, and solves it with a matrix-free GMRES/QTT scheme. Combined with nonequilibrium DMFT (second-order Born impurity self-energy) for an interaction quench of the half-filled Hubbard model on the Bethe lattice, the method shows that vertex corrections reverse the short-time sign of the order-parameter fluctuation σ_m and that both the peak fluctuation and its decay time grow as the quench approaches the nonthermal critical point.","tokens_in":25507,"tokens_out":1117,"duration_ms":13740,"significance":"Computing vertex-corrected nonequilibrium two-particle correlators is a recognized bottleneck for interpreting time-resolved Raman and RIXS experiments. The linear equation for G′ together with the matrix-free QTT Krylov solver is a concrete, reusable algorithmic advance: it reduces memory from O(N_t^4) to two-time operations (or O(D_max^3) in QTT form) and extends naturally to denser kernels via Jacobian-vector products (End Matter). Short-time benchmarks against direct GCH pump-probe simulations (End Matter Fig. 4), public code (tensor4all-rs), and a controlled DMFT demonstration that vertex corrections are qualitatively essential strengthen the contribution. The physical finding that nonthermal criticality is visible in two-particle fluctuations is of independent interest within the nonequilibrium DMFT community.","major_comments":[{"comment":"The central methodological claim (Eq. 3 and the matrix-free solver) is sound and well supported by the short-time GCH pump-probe benchmark (End Matter Fig. 4). No load-bearing technical error was found in the derivation or the QTT/GMRES implementation.","section":null},{"comment":"Figs. 2(e,f) and 3(b): the claim that fluctuation decay time τ_χ grows toward the nonthermal critical point U_nc_c rests on a narrow window of U_f (1.1–1.3) and a fit interval t∈[24,32) that coincides with the stated GMRES limit t_max=32. The text itself notes that the vertex-corrected τ_χ may appear to approach a different critical point and that longer times are needed. The abstract and conclusion should qualify this statement more carefully (e.g., “within the accessible window, both σ_m^max and τ_χ increase as U_f approaches U_nc_c”) so that the strongest physical claim is not overstated relative to the numerical reach.","section":null}],"minor_comments":[{"comment":"Fig. 1(c) caption and main text: the equal-time fluctuation is written σ_m(t_p)=−Im χ^<(t_p,t_p); a one-sentence reminder that this equals N(⟨m̂ m̂⟩−⟨m̂⟩²) (Eq. 6) would help readers who jump to the figure.","section":null},{"comment":"End Matter / Algorithm 1: state the restarted GMRES subspace size m and the typical number of outer restarts used for the production runs in Figs. 2–3, so that the reported residual tolerances are reproducible.","section":null},{"comment":"Supplemental Material Eqs. (S10)–(S11): the sparse structure of K for the second-order self-energy is useful; a brief cross-reference from the End Matter paragraph on computational cost would make the contrast with dense kernels (e.g., GW) clearer.","section":null},{"comment":"Fig. 3(a): the linear fits used to extract U_nc_c from 1/τ and ω_H should report the fit ranges and uncertainties (or at least the numerical value of U_nc_c) in the caption or Supplemental Material.","section":null},{"comment":"Notation: the same symbol χ is used for the full contour correlator and for its lesser component; a consistent superscript convention (already used in places) would reduce ambiguity.","section":null},{"comment":"References: the recent functional-derivative Raman work on the Falicov–Kimball model (Ref. 36) is appropriately cited; a short sentence distinguishing the present linear-equation approach from that finite-difference GCH pump-probe scheme would help non-specialists.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The paper is a solid methods-plus-application contribution suitable for a high-impact condensed-matter journal. The second-order Born truncation is a limitation of the demonstration, not of the method, and is already standard in the nonthermal-criticality literature the authors build on. I would not require a non-perturbative impurity solver for acceptance. The only substantive editorial request is a more cautious wording of the fluctuation-criticality claim given t_max=32."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The real news is the method. They take the generalized Keldysh contour with a branch-dependent virtual probe, differentiate the Dyson equation once, and get a linear integral equation for G' that never builds the four-time kernel K. They then solve it matrix-free with QTT-GMRES. That combination is new and directly attacks the memory wall that has kept most nonequilibrium BSE work stuck at the bubble level.\n\nWhat they do well is keep the derivation clean, control the numerics (short-time agreement with direct GCH pump-probe, residual and truncation tolerances stated), and ship public libraries. The physics application is also useful: after an interaction quench in second-order DMFT, the vertex correction flips the short-time sign of the order-parameter fluctuation σ_m, and both the peak height and the decay time grow as one approaches the nonthermal critical point. That is a concrete qualitative message that people working on nonthermal criticality and time-resolved spectroscopies will want.\n\nSoft spots are real but secondary. The impurity self-energy is only second-order Born, so the precise location of U_nc and the size of the vertex effect could shift with a better solver; the authors already flag this and show how the same linear equation works for denser kernels (e.g., GW via JVP). Accessible time is limited to t_max=32 by GMRES convergence; they note that divide-and-conquer QTT ideas should help. Neither issue undercuts the methodological claim.\n\nThis is for people who actually compute nonequilibrium spectra or fluctuations. It deserves a serious referee and is ready for the community to use and extend. I would engage.","headline":"Solid computational advance: a matrix-free linear equation for vertex-corrected nonequilibrium two-particle functions, with a clean demonstration that vertices reverse short-time fluctuation dynamics near a nonthermal critical point.","tokens_in":26107,"tokens_out":441,"would_cite":true,"duration_ms":4660,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Vertex corrections reverse short-time order-parameter fluctuation dynamics after a quench and grow both amplitude and lifetime near a nonthermal critical point.","keywords":["nonequilibrium Green’s functions","generalized Keldysh formalism","vertex corrections","quantics tensor trains","nonequilibrium DMFT","order-parameter fluctuations","nonthermal critical point","interaction quench"],"falsifier":"Repeat the same interaction quench with a higher-order or non-perturbative impurity solver and check whether the short-time sign reversal of the fluctuation and the growth of peak amplitude and decay time near the nonthermal critical point survive.","tokens_in":26110,"feed_emoji":"⚡","tokens_out":941,"duration_ms":8888,"temperature":0.7,"pith_summary":"Time-resolved spectroscopies now measure collective fluctuations in driven quantum materials, but theory has struggled to compute the needed two-particle correlation functions once vertex corrections (collective effects) are kept. This paper introduces a contour-dependent virtual probe field in a generalized Keldysh setup and converts the response into a single linear integral equation for the variation of the one-particle Green’s function, solved without building the four-time vertex kernel by a matrix-free Krylov method on quantics tensor trains. Combined with nonequilibrium dynamical mean-field theory for an interaction quench of the antiferromagnetic Hubbard model, the method shows that including vertex corrections qualitatively reverses the early-time evolution of the order-parameter fluctuation and that both the peak fluctuation and its decay time increase as the system approaches the nonthermal critical point. The result supplies a practical route to the correlation functions that experiments and entanglement diagnostics are beginning to demand far from equilibrium.","feed_headline":"Vertex corrections flip quench fluctuation dynamics","feed_subtitle":"A linear probe equation plus tensor trains shows amplitude and lifetime grow near nonthermal criticality","key_machinery":"The linear integral equation for the Green’s-function variation under a contour-dependent virtual probe (Eq. 3), solved by a matrix-free GMRES Krylov method that only needs the two-time action of the vertex kernel and represents two-time objects as quantics tensor trains.","core_discovery":"Once the unperturbed one-particle Green’s function is known, the full vertex-corrected nonequilibrium two-particle correlation function is obtained from the solution of a linear integral equation for its response to a contour-dependent virtual probe; applied to a quench into a nonequilibrium symmetry-broken state, that response shows that vertex corrections reverse the short-time fluctuation dynamics and that both the fluctuation amplitude and its decay time grow near the nonthermal critical point.","pith_inferences":["If the nonthermal critical enhancement of fluctuations is generic, pump-probe experiments that track equal-time spin or charge variance should see a clear peak in both amplitude and recovery time as the drive strength is tuned through the nonthermal critical region.","Encoding the probe time as an extra quantics leg, as the authors sketch, would turn a sequence of independent linear solves into one compressed solve and could make dense frequency-domain spectra routine.","The same contour-probe construction applies immediately to other one-particle operators (current, pairing, orbital occupancy), so the method can target the full set of two-particle channels that enter quantum Fisher information and related entanglement witnesses."],"forward_implications":["Nonequilibrium spectra measured by time-resolved Raman or RIXS can be computed with the vertex corrections that control collective modes, rather than with the bubble approximation alone.","Order-parameter fluctuation amplitude and lifetime become practical diagnostics of nonthermal criticality, complementary to the order parameter itself.","The same linear-response equation can be fed any self-energy whose Jacobian-vector product is available, including diagrammatic or quantum-Monte-Carlo impurity solvers.","Quantics-tensor-train compression makes the nine-component generalized Keldysh storage feasible for multi-orbital and larger-lattice systems."],"fun_headline_variants":["Vertex corrections reverse short-time quench fluctuations","Virtual probe flips nonequilibrium order-parameter dynamics","Fluctuation amplitude and lifetime grow near nonthermal criticality","Linear probe equation yields full vertex-corrected correlations","Tensor-train solver shows vertex role in quench fluctuations"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The impurity self-energy is kept only at second-order Born level (Hartree–Fock plus second-order diagrams); a different truncation or a non-perturbative impurity solver could move the nonthermal critical point and change the size of the vertex corrections.","fun_headline_variants_meta":{"raw":{"variants":["Vertex corrections reverse short-time quench fluctuations","Virtual probe flips nonequilibrium order-parameter dynamics","Fluctuation amplitude and lifetime grow near nonthermal criticality","Linear probe equation yields full vertex-corrected correlations","Tensor-train solver shows vertex role in quench fluctuations"]},"model":"grok-4.5","effort":"low","cost_usd":0.005614,"raw_usage":{"total_tokens":1459,"prompt_tokens":737,"num_sources_used":0,"completion_tokens":77,"cost_in_usd_ticks":56140000,"prompt_tokens_details":{"text_tokens":737,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":645,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":737,"tokens_out":77,"duration_ms":6270,"temperature":1.0,"reasoning_tokens":645,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-14T07:21:39.309568+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Repeat the same interaction quench with a higher-order or non-perturbative impurity solver and check whether the short-time sign reversal of the fluctuation and the growth of peak amplitude and decay time near the nonthermal critical point survive.","supporting_citations":[],"review_version":1}