{"id":"10610029-6926-4bd4-af36-01adfa7fa15d","arxiv_id":"2602.21541","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Exact multi-particle propagator for the 1D Hubbard model, derived from the nested Bethe ansatz as a contour integral — the first Yudson representation for a model with spin.","lead":"This paper derives an exact integral formula for the propagator of the one-dimensional Hubbard model, expressing exactly how N fermions with spin evolve on an infinite line. It is the first such 'Yudson' representation for an integrable model with spin degrees of freedom, and gives an analytic starting point for nonequilibrium and open-system calculations beyond the reach of numerics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The t=0 reduction of Theorem 1 rests on an asserted braid-algebra cancellation (eqs. (59)-(63)) in Lemma 2; identity (50) for n≥3 is not fully demonstrated and would invalidate the initial condition if false.","rationale":"The paper's proof has two parts. Step (a) is solid: the integrand is a superposition of nested Bethe wave functions, which solve the Schrödinger equation; the contours are compact and avoid singularities, so differentiation under the integral is unproblematic. The only place where a hidden error could enter is the t=0 reduction. Lemma 1 is supported by Proposition 1, whose proof uses residue calculus with a cancellation analogous to (43); the reader found the M=1 base case correct, and the scalar nature of the spin amplitudes makes that part less delicate. Lemma 2 is more delicate because the Y-operators are noncommuting matrices. The critical line is the cancellation of the 1/(s_l-s_m) poles in the z_m integration. Identity (63) is dimension-dependent (three-site basis) and the paper does not display the full algebra for arbitrary permutation words; a reader cannot verify from the text that the 'triangle' factor (62) always appears with all other factors regular. This is exactly the step that, if wrong, breaks (50) and hence the initial condition. The external consistency checks (N=2,M=1 prior result, Tracy–Widom limit) do not exercise the n≥3 induction, so they cannot rule out a failure there. The proposed symbolic test directly computes the contested identity for the smallest case where n≥3 appears, without using the proof's asserted cancellation. If it reproduces the determinant for generic parameters and spin sectors, it would strongly support the proof; if not, the theorem is false. Thus the appropriate verdict remains CONDITIONAL pending this check.","tokens_in":23995,"tokens_out":20986,"duration_ms":170917,"concrete_test":"Symbolically evaluate the left-hand side of (25) for N=4, M=2 with generic ordered positions y and x (e.g., y=(0,1,2,3), x=(0,1,2,3)) and spin configurations a,b, by computing the z-integrals as exact residues using the explicit word decomposition (51) for each P — without invoking identity (63) as a simplification. Check that the sum over P∈S_N^{(n)} for n=3,4 vanishes and the total equals det[δ_{x_j,y_k}δ_{a_j,b_k}]. If any residual pole 1/(s_l-s_m) survives or the determinant is not reproduced, Lemma 2 and Theorem 1 are falsified; if it passes, the braid-cancellation works in the first nontrivial case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is Theorem 1, proved by showing the RHS satisfies both the Schrödinger equation and the initial condition. Step (a) is immediate; the load-bearing step is Step (b), specifically Lemma 2's identity (50) for P∈S_N^{(n)}, n≥3. After the z_n and z_m residue integrations, the sum over P is expressed as T_c(l,m;P) in (64). The proof must show that the apparent poles of (59) at s_m=s_l+2iu cancel. The paper asserts that by the Yang–Baxter relation one can choose a decomposition of P so that the triangle factor (62) appears, and that identity (63), verified by 'direct evaluation' on three-site states, makes the residue regular at s_m=s_l+2iu. This is the least-secure element of the proof: the decomposition is not constructed for general P (Appendix C gives a derivation but with a 'direct calculation' step), and (63) is an operator identity whose application to the full residue chain requires all other factors to be regular — a condition that is asserted, not proven. If (50) fails for any N≥3, the initial condition (7) is not satisfied and the propagator formula is false. The reader verified (63) on three-site states and the N=2 case, but the general induction step remains unverified. This is an internal-consistency issue, not a dispute with existing results.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims an exact multiple-contour integral formula for the N-particle propagator of the one-dimensional Fermi–Hubbard model on an infinite lattice, in the sector with M down spins, for arbitrary complex interaction u≠0. Theorem 1 (eq. (21)) expresses the propagator as N charge contour integrals over circles |z_j|=r^{N−j} and M spin contour integrals over small contours Γ_s around {s_j+iu}, applied to the nested Bethe wave function φ(x;a|z;λ). The proof verifies that the right-hand side satisfies the Schrödinger equation and, at t=0, reduces to the determinant initial condition (7). The t=0 reduction is carried out through Lemma 1 (spin amplitudes expressed by Y-operators) and Lemma 2 (charge integrations reproduce the determinant). The paper also applies the formula to a dephased tight-binding chain and to the two-body-loss Hubbard model via GKSL duality, and notes a homogeneous-limit check against the Tracy–Widom ASEP identity.","tokens_in":24259,"tokens_out":18924,"duration_ms":155220,"significance":"If correct, this is a significant technical advance: it provides the first Yudson-type integral representation for the Hubbard model, does not rely on the string hypothesis, and gives a concrete route to exact finite-particle nonequilibrium dynamics, including non-Hermitian and open-system applications. The proof strategy is sound in principle: it is a direct verification of the Schrödinger equation and the initial condition, rather than an assumed completeness of Bethe states. The paper also contains useful external benchmarks: the N=2, M=1 case reproduces earlier work, and the homogeneous limit reduces to the known Tracy–Widom ASEP identity. The main risk is the t=0 reduction, which depends on nontrivial braid/Y-operator cancellations; the manuscript does not yet make this part fully transparent or fully rigorous.","major_comments":[{"comment":"The proof of identity (50) for n≥3 hinges on the assertion that, after the z_n and z_m residue integrations, the apparent pole of factor (59) at s_m=s_l+2iu cancels because a decomposition of P can be chosen so that the triangle factor (62) appears and the operator identity (63) holds. This is load-bearing: if the cancellation fails, the t=0 limit does not reduce to the determinant (7). However, the decomposition is not constructed for general P; Appendix C.2 treats only l<m<n with P^{-1}(m)<P^{-1}(l), and the decisive step is stated as a 'direct calculation' without displaying the calculation. The regularity of all remaining factors at s_m=s_l+2iu is asserted, not proved. The operator identity (63) is likewise said to be verifiable on three-site states, but the verification is not shown. I recommend turning this into a complete lemma: explicit construction of the decomposition, proof of","section":"§4.2, eqs. (59)–(63), Appendix C.2"},{"comment":"The cancellation T_c(l,m;P)+T_c(l,m;Π_{l,m}P)=0 is central to proving (50) and hence the initial condition. The proof of (66) in Appendix C.3 assumes 'without loss of generality' μ<ν and performs a one-line manipulation involving an ordered product of Y-operators. This step is not fully algebraic: it does not spell out how the decomposition of Π_{l,m}P is obtained from that of P, nor why the residues commute with the insertion of Y(0)=I and with the unitarity relation Y(λ)Y(−λ)=I. Since the whole cancellation argument depends on exact operator orderings, I ask for a fully detailed derivation, either algebraic or diagrammatic with all orderings explicit, rather than the schematic figure 4.","section":"§4.2, eqs. (64)–(66), Appendix C.3"}],"minor_comments":[{"comment":"The induction over M is compressed. In the passage after the λ_1 integration for R∈S_M^{(1)}, the relabeling of α_j, β_j and λ_2,...,λ_M that allows application of the induction hypothesis is not displayed. Please write out this reduction explicitly.","section":"§4.1, Proposition 1"},{"comment":"The statement that the 'omitted part' is holomorphic inside the z_1-contour relies on the chosen radii r^{N-j}; a short justification showing that the Y-operator denominators cannot vanish inside the contour would improve readability.","section":"§4.2, below (53)"},{"comment":"The homogeneous-limit check is valuable but the displayed derivation of (29) from (31) is terse. In particular, the sign in the scattering factor after the change of variables ξ_j=(λ_j−iu)/(λ_j+iu) is not obvious; please expand the computation.","section":"Remark 5"},{"comment":"No numerical or exact-diagonalization check for N≥3 is provided. Given the complexity of Lemma 2, a small benchmark (e.g., N=3, M=1 or M=2 on a finite lattice) would substantially increase confidence in the intricate residue cancellations.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the claimed result is significant. I see no circularity problem: the nested Bethe wave function is imported from the literature, and the homogeneous limit provides a nontrivial external benchmark. My main concern is the completeness of the proof of Lemma 2, especially the braid/Y-operator cancellations leading to (50). The missing details appear fillable, but they are load-bearing and should be supplied before publication. I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. This is the first Yudson-type propagator for the Hubbard model at general N and M, and I think the result is new and worth taking seriously. The formula in Theorem 1 is a multiple contour integral of the nested Bethe wave function; it goes beyond the existing Yudson representations for Lieb-Liniger, XXZ, Gaudin-Yang, and beyond the authors' own N=2, M=1 case with y1=y2. They prove it by verifying the Schrödinger equation and the initial condition, which is the right strategy.\n\nWhat's good: the construction is honest. The M=1 base case checks out; the homogeneous limit reproduces the Tracy–Widom ASEP identity; and the N=2, M=1 case matches their earlier paper. The Bethe wave function is imported from standard sources, so there's no circularity. They also spell out the two open-system applications. The notation is heavy, but the graphical explanation of the Y-operator structure helps.\n\nThe soft spot is real but not fatal. The load-bearing step is Lemma 2, identity (50) for n≥3, where the sum over permutations with P(1)=n must vanish. The proof says that by the Yang–Baxter relation one can choose a decomposition of P so that a triangle factor appears, and that identity (63) makes the residue regular. But the decomposition is not constructed for general P, and (63) is verified only on three-site states by 'direct evaluation.' The reader spot-checked several components and found nothing wrong, but the general induction step is not fully displayed. If that cancellation failed, the initial condition would break, so this is the part a referee should scrutinize. There are also two smaller assumptions: the non-normalizable Bethe states on the infinite lattice are used as a complete set (Remark 1), which is reasonable but not argued, and the distinctness condition behind some exponent inequalities is left implicit.\n\nOn balance, I think the theorem is likely correct. The external consistency checks and the spot-checks give me moderate confidence. The proof gap is more in the presentation than in the underlying idea. The main audience is people working on exact nonequilibrium dynamics of integrable systems and on open quantum models with complex interactions. This paper deserves a serious referee rather than a desk reject. Send it to peer review, with instructions to verify Lemma 2 carefully and to ask for a more explicit proof of the braid cancellation.","headline":"First Yudson-type propagator for the Hubbard model at general N,M; a solid, new result whose proof has one load-bearing braid-cancellation step that needs a careful referee.","tokens_in":24874,"tokens_out":4093,"would_cite":true,"duration_ms":36585,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B23"],"pacs":["71.10.Fd","05.30.-d"],"model":"deepseek-v4-flash","headline":"The multi-particle propagator of the one-dimensional Hubbard model equals an explicit multiple contour integral of nested Bethe wave functions, exact for all particle numbers, times, and complex interactions.","keywords":["Hubbard model","nested Bethe ansatz","propagator","contour integral","exact time evolution","open quantum systems","dephasing noise","two-body loss"],"falsifier":"For N=3, M=1, evaluate the right side of (21) at t=0 by numerical contour integration or by exact residue sums for distinct initial and final configurations with y_1−x_2≥0 and compare with the determinant δ_{x1,y1}δ_{x2,y2}δ_{x3,y3}δ_{a1,b1}δ_{a2,b2}δ_{a3,b3}; a nonzero result for any permutation P∈S_3^(3) would falsify Lemma 2 and hence Theorem 1.","tokens_in":23805,"feed_emoji":"⚛️","tokens_out":5380,"duration_ms":53000,"temperature":0.7,"pith_summary":"The paper aims to prove that the full time evolution of any finite number of fermions in the one-dimensional Hubbard model can be written as a single explicit multiple contour integral, with no quantization conditions such as the string hypothesis. If true, any finite-particle initial wave function can be evolved exactly by plugging into that integral, which has so far been possible only for simpler integrable models. The formula also holds for complex interaction strengths, which is what makes it directly applicable to open quantum systems such as a tight-binding chain with dephasing noise and the Hubbard model with two-body loss. The authors show the contour integral solves the Schrödinger equation and, via nested residue calculus, that it returns the correct delta-function initial state at t=0.","feed_headline":"One contour integral solves Hubbard-model time evolution","feed_subtitle":"Any particle number, any complex interaction: exact dynamics for the Hubbard model and its dissipative cousins.","key_machinery":"The load-bearing object is the nested Bethe wave function (13)–(14), in which charge-plane waves are dressed by spin scattering amplitudes that are themselves Bethe states of an inhomogeneous XXX spin chain. Two structural identities carry the proof: the Y-operator relation (19), expressing scattering amplitudes for arbitrary permutations through adjacent transpositions, and the (braid) Yang–Baxter relation (26), which makes those expressions well-defined. The initial-condition check then proceeds by nested induction: spin-contour residues prove A_id=δ_{a,b}, and charge-contour residues with a graphical bookkeeping of Y-operators cancel all higher permutations, leaving the determinant (7).","core_discovery":"For the N-particle, M-down-spin sector of the infinite-lattice Fermi–Hubbard model, the propagator ψ_t(x;a|y;b) is given by the multiple contour integral (21): product over charge rapidities z_j on small circles |z_j|=r^{N−j}, product over spin rapidities λ_k on contours enclosing s_{β_k}+iu, with the nested Bethe wave function φ(x;a|z;λ) and energy factor e^{−iE(z)t}. Each nested scattering amplitude is itself a Bethe wave function of an inhomogeneous XXX spin chain. The proof verifies the two defining properties of the propagator: the Schrödinger equation follows from the nested Bethe ansatz, and the initial condition is recovered by successive residues in spin and then charge rapidities,","pith_inferences":["The residue-induction structure (spin first, then charge) looks robust enough to survive extension to the SU(n) generalization of the Hubbard model mentioned in the outlook; the same Y-operator/Yang–Baxter machinery is expected to carry over.","If the thermodynamic limit of the formula can be taken for density-matrix elements, it would give exact finite-density transport coefficients for the dephased chain, going beyond the finite-particle sector treated here.","The complex-interaction version suggests a direct route to the distribution of particle-loss events in the dissipative Hubbard model: because the propagator encodes the no-jump evolution, counting statistics of loss should follow by summing over loss times.","At large times and for step initial states, the exact integral should reproduce the known diffusive/KPZ-type scaling of the dephased tight-binding chain; recovering that scaling from the formula would validate its use in the hydrodynamic regime."],"forward_implications":["Any N-particle state can be evolved exactly by inserting the integral formula into the expansion (5)-(6), giving a microscopic handle on nonequilibrium dynamics of the Hubbard model.","Because the formula holds for complex u∈C∖{0}, it yields, through the appendix-A duality, exact density-matrix elements for the tight-binding chain with dephasing noise and loss probabilities for the Hubbard model with two-body loss.","The derivation bypasses the string hypothesis; no assumptions about bound-state rapidity patterns are needed.","The N=2, M=1 case has already given exact density profiles for dephased domain-wall states; the general formula provides the basis for the full counting statistics of the time-integrated current.","In the homogeneous limit s_j=0, the spin part of the identity reduces to a known exclusion-process integral identity, placing the result in the family of exact propagator integral representations known for other integrable systems."],"fun_headline_variants":["Exact Hubbard propagator via nested Bethe ansatz","Contour integrals give exact Hubbard time evolution","No string hypothesis: exact Hubbard dynamics formula","Arbitrary particle count: exact Hubbard propagator","Hubbard propagator from nested Bethe integrals"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The initial-condition proof depends on a pole-cancellation identity (63) supported by a direct evaluation on local basis states, and on the assertion that for every permutation one can choose an adjacent-transposition decomposition making the cancellation appear; if that braid-cancellation fails for some particle number, the t=0 integral would not equal the required determinant.","fun_headline_variants_meta":{"raw":{"variants":["Exact Hubbard propagator via nested Bethe ansatz","Contour integrals give exact Hubbard time evolution","No string hypothesis: exact Hubbard dynamics formula","Arbitrary particle count: exact Hubbard propagator","Hubbard propagator from nested Bethe integrals"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00107,"raw_usage":{"total_tokens":4254,"prompt_tokens":612,"completion_tokens":3642,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":356,"completion_tokens_details":{"reasoning_tokens":3569}},"tokens_in":356,"tokens_out":3642,"duration_ms":24649,"temperature":1.0,"reasoning_tokens":3569,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T21:03:19.618981+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For N=3, M=1, evaluate the right side of (21) at t=0 by numerical contour integration or by exact residue sums for distinct initial and final configurations with y_1−x_2≥0 and compare with the determinant δ_{x1,y1}δ_{x2,y2}δ_{x3,y3}δ_{a1,b1}δ_{a2,b2}δ_{a3,b3}; a nonzero result for any permutation P∈S_3^(3) would falsify Lemma 2 and hence Theorem 1.","supporting_citations":[],"review_version":1}