{"id":"9d403326-0350-4aed-a654-15e631c781de","arxiv_id":"1908.05578","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Time-dependent defects described by Bäcklund transformations preserve integrability for NLS, KdV, and mKdV, and the moving-defect KdV system admits peaked (peakon) solutions.","lead":"This paper studies integrable wave equations whose internal 'defect' (an impurity) moves with time, and shows the systems still have infinitely many conservation laws. It also finds that a KdV soliton can have a sharp peak (a peakon) when the moving defect sits at the soliton's crest, which is a new way to get peakons outside the Camassa-Holm family.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (4.24) is asserted, not proved: with a moving boundary, the equal-space r-matrix relation for the piecewise transition matrix may acquire c'(t) terms, so Liouville integrability of the defect NLS system is not established.","rationale":"The reader's weakest assumption identifies the same point: (4.24) is load-bearing for the Liouville half of the integrability claim. I agree. Proposition 1's conservation-law proof is explicit and accounts correctly for c'(t) through (4.8)-(4.10); I do not see a fatal error there. The Lagrangian claims are lengthy but internally consistent. The peakon construction also works for appropriately chosen β, though the paper should state the reality condition β²≥8k² sech²γ and clarify the signs/cases in (6.3); this is a minor omission, not the main threat. The decisive gap is the uncomputed r-matrix bracket at the moving boundary. Because the proof is absent rather than known wrong, the correct verdict is conditional, not reject: if a direct computation confirms (4.24), the paper's central claim stands; if it produces boundary terms, the Liouville-integrability statement needs correction.","tokens_in":15198,"tokens_out":13434,"duration_ms":138452,"concrete_test":"Recompute the left side of (4.24) directly from (4.23) and (4.12): write {M_1(x,t,λ), M_2(x,t,μ)} for t,μ arbitrary, using the BT identities (2.19) to relate u and ˜u at x=c(τ), and keep all boundary terms arising from differentiating the branch switch c(τ)=x. Verify whether the result equals [r(λ−μ), M(x,t,λ)⊗M(x,t,μ)] with no additional c'(t)δ-terms. A companion check is to build the x-ordered monodromy with the defect matrix B evaluated at c(t), compute its equal-time bracket, and test whether its trace is stationary; this would settle whether the omitted moving-boundary computation hides extra terms.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central integrability claim has two parts: infinite conservation laws (Proposition 1, proved via (4.1)-(4.11)) and Liouville integrability via the r-matrix (Section 4.2). The load-bearing weak point is the latter. In (4.23), M(x,t,λ) is defined by switching between time-ordered exponentials of V(u) and V(˜u) according to x<c(t) or x≥c(t), and (4.24) is then said to follow 'immediately' from the canonical property of the BT. That inference is not shown. Each branch separately satisfies (4.16), but the equal-space Poisson bracket of the piecewise object is a Poisson bracket in t at fixed x; when the boundary c(t) sweeps past a fixed x, the branch in the time-ordered integrand switches at τ0(x), and differentiating the switching point can produce boundary terms proportional to c'(t). The canonical transformation statement (4.19)-(4.22) concerns a Pfaffian form integrated over all t and does not, by itself, control the r-matrix bracket of these time-ordered exponentials. In addition, M_T and ˜M_T are formal unless u and ˜u are extended to the whole (x,t) plane, since each time-ordered exponential needs the potential at times when x is on the other side of the defect. If (4.24) fails, the involution of the trace integrals is unproved; only the conservation-law half of integrability remains rigorous.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies (1+1)-dimensional integrable soliton equations in the AKNS class with a defect located at a time-dependent position x=c(t). The defect condition is taken to be a Bäcklund transformation evaluated at the moving point x=c(t). The authors claim three principal results: (i) such defect systems possess infinitely many conserved quantities, with the generating function I(λ) in Eq. (4.1) proved time-independent; (ii) the defect NLS, KdV, and mKdV systems admit Lagrangian descriptions, and Liouville integrability can be established via the classical r-matrix method; (iii) the defect KdV equation with a defect moving at constant speed admits explicit peakon solutions, including a multi-peakon solution for multiple defects. The conservation-law proof (Proposition 1) is essentially complete and correct, while the r-matrix argument in Section 4.2 is only asserted.","tokens_in":15528,"tokens_out":7429,"duration_ms":69520,"significance":"If the r-matrix gap were closed, this would be a natural extension of the fixed-defect results by Caudrelier and by Caudrelier–Kundu, and the time-dependent-defect peakon solutions would be a genuinely new phenomenon for KdV-type equations. The paper's proofs are partly machine-checkable in the sense of being direct symbolic computations: the conservation-law derivation in Eqs. (4.5)–(4.10) is explicit and correct, and the peakon solution in Proposition 3 is an explicit verification. The central weakness is that the Liouville-integrability claim rests on an unproved assertion about the equal-space Poisson bracket of a piecewise time-ordered transition matrix.","major_comments":[{"comment":"The statement that \"due to the canonical property of the transformation, we immediately conclude\" that the piecewise transition matrix M(x,t,λ) satisfies the same r-matrix relation is not justified. For a fixed x, the branch of M in (4.23) switches at a time τ0 satisfying c(τ0)=x, so the time-ordered exponentials (4.15) are integrals whose local branch changes as τ crosses τ0. When computing the equal-space Poisson bracket (4.12) at fixed x, derivatives of the switching time with respect to t can introduce boundary terms proportional to c'(τ0) in addition to the bulk r-matrix terms. The canonical-transformation statement (4.18)–(4.22), which concerns invariance of a Pfaffian form with a generator W, does not by itself control the r-matrix bracket of these time-ordered products. An explicit computation is needed; without it, the involution of the trace integrals and the claimed Liouville integrability of (3.1) are not established.","section":"§4.2, Eq. (4.24)"},{"comment":"The matrices M_T(x,t,λ) and \\tilde M_T(x,t,λ) are defined as time-ordered exponentials over all τ∈(-∞,t) of V(u(x,τ)) and V(\\tilde u(x,τ)), respectively. However, u is a priori defined only for x>c(t), and \\tilde u only for x<c(t). For a fixed x, there are generally times τ at which x lies on the opposite side of c(τ), so the integrand in the time-ordered exponential is not defined. The manuscript should state the additional hypotheses needed (for example, monotonicity of c(t) combined with smooth extensions of the fields across the defect) or redefine the transition matrix so that the time integration respects the moving boundary. This gap affects both (4.24) and the formal definition of the monodromy.","section":"§4.2, Eqs. (4.15) and (4.23)"},{"comment":"The assertion that the multi-defect transition matrix satisfies the same r-matrix relation is made without proof and inherits the gap already present in (4.24). In addition, the multi-defect setting requires that the ordering c1(t)<c2(t)<...<cn(t) be preserved for all times, and the switching times for each defect as t varies interact nontrivially with the equal-space Poisson bracket. Please provide the computation or state the restrictions on cj(t) that make the argument valid.","section":"§5, Eq. (5.5)"}],"minor_comments":[{"comment":"The Lagrangian derivations are only sketched with the phrase \"after some algebra.\" Since these claims are not used in the subsequent integrability proof, I do not view this as blocking, but the paper would be more self-contained if at least one reduction (e.g., from (3.18c)–(3.18d) to (3.1c)–(3.1d)) were shown explicitly.","section":"§3.2, Claims 1–3"},{"comment":"The solution (6.5) has a corner at ξ=0, so the quantities u_x and \\tilde u_x in the defect conditions (3.2c)–(3.2d) are not defined in the usual classical sense at the defect point. The paper should state explicitly that the defect conditions are interpreted with one-sided limits at x=c(t).","section":"§6, Eq. (6.5)"},{"comment":"There are several copy-editing slips, e.g., \"d efect\" in the abstract, \"A n interesting\" in the Introduction, and inconsistent spacing in \"B¨ acklund.\" These should be corrected but do not affect the mathematics.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"To the editor: The paper contains a correct and clean proof of infinitely many conservation laws for moving defects and an explicit, verifiable peakon solution. The main unresolved point is the r-matrix step in Section 4.2, which is load-bearing for the Liouville-integrability claim. I believe this is fixable if the authors carry out the Poisson-bracket computation for the piecewise transition matrix or, failing that, restrict to monotone c(t) and global field extensions. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline first: the paper does one solid thing and one half-proved thing. The solid thing is Proposition 1, the conservation laws for defects moving with x=c(t); the proof is complete and I could not find a gap. The half-proved thing is Section 4.2, where Liouville integrability via the r-matrix is asserted rather than derived, and the moving boundary makes that assertion non-obvious. The peaked-soliton observation is a nice side result and it checks out.\n\nWhat is new: the defect is at a time-dependent position rather than a fixed one, and the defect condition is a Bäcklund transformation evaluated at x=c(t). The paper then gives Lagrangian descriptions for NLS, KdV and mKdV with moving defects, and proves an infinite family of conserved quantities. The Lagrangian claims are plausible but the computations are mostly suppressed ('after some algebra'); that is a minor weakness, not a fatal one, since the variation setup is shown. The multiple-defect generalization is straightforward. Citation pattern looks right: earlier fixed-defect work is credited.\n\nThe real soft spot is equation (4.24). The paper jumps from 'the transformation is canonical' to 'M satisfies the same r-matrix relation.' That is not immediate. M is piecewise in x, switching branches along c(t), and the equal-space Poisson bracket is taken at fixed x; when t crosses the value where c(t)=x, the time-ordered exponential switches from V to \\tilde V. Boundary terms involving c'(t) could enter. The canonical-transformation statement (4.19)-(4.22) is about a Pfaffian form integrated over all t and does not control the bracket of these time-ordered exponentials. So the Liouville integrability claim is not established as written. I would treat this as a load-bearing gap in the paper's advertised scope, though it is probably fixable.\n\nThe peakon (Proposition 3) is an explicit ansatz. I checked the two defect conditions: with alpha2=alpha1^{-1}, the left side of (6.3) cancels and the right side vanishes because the two denominators are equal, so no hidden parameter relation is needed; you just need beta large enough for the square root to be real. The authors are careful that this is a corner at the moving defect, not a weak peakon in the Camassa-Holm sense.\n\nWho gets value: people working on integrable defects or Bäcklund transformations. The paper deserves a serious referee, but the referee should insist that Section 4.2 either supply the missing computation or be downgraded to a conjecture.","headline":"Genuine extension to moving integrable defects, with a solid conservation-law proof and a load-bearing r-matrix step that is asserted rather than shown.","tokens_in":16027,"tokens_out":5153,"would_cite":false,"duration_ms":52252,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37K10","35Q53","35Q55","37K35"],"pacs":[],"model":"deepseek-v4-flash","headline":"Time-dependent defects defined as Bäcklund transformations preserve integrability of AKNS-class soliton equations, and the defect KdV equation admits peaked soliton solutions.","keywords":["integrable defect","time-dependent defect","Bäcklund transformation","AKNS system","conservation laws","classical r-matrix","peakon","KdV equation"],"falsifier":"Compute the equal-space Poisson bracket {M1(x,t,λ), M2(x,t,μ)} explicitly for the piecewise-defined transition matrix near the moving boundary x=c(t) with c'(t)≠0. If the result differs from [r(λ−μ), M(x,t,λ)⊗M(x,t,μ)] by terms supported at the boundary, the Liouville-integrability claim for the time-dependent defect NLS system collapses. Short of that analytic check, a numerical test of the first few conserved quantities for the defect KdV equation with a non-constant c(t) would provide evidence.","tokens_in":15012,"feed_emoji":"🌊","tokens_out":7239,"duration_ms":66698,"temperature":0.7,"pith_summary":"This paper asks whether a defect that moves, at x=c(t), can be inserted into a classical integrable soliton equation without destroying exact solvability. The authors define the defect condition as a Bäcklund transformation evaluated at the moving point rather than over the full line, and argue that the answer is yes: NLS, KdV, and mKdV with such defects possess infinitely many conserved quantities and are Liouville integrable. The same construction yields Lagrangian descriptions, and in the KdV case a defect moving at the soliton speed $4k^{2}$ produces a peaked wave u(x,t)=$2k^{2}$ $sech^{2}$(|k(x-$4k^{2}$t)|+γ), a type of solution normally associated with shallow-water peakon equations rather than KdV.","feed_headline":"Moving defects keep soliton equations integrable","feed_subtitle":"A Bäcklund-defect at x=c(t) yields conserved quantities and peaked KdV waves.","key_machinery":"The engine is the Bäcklund transformation used as an internal boundary condition: at x=c(t) the fields on the left and right are related by the Darboux matrix B, so the gauge connection between the two Lax pairs is enforced only at the moving point. This makes the ratio B11+B12Γ at the defect the quantity whose logarithm cancels the boundary flux in the conservation-law proof, and the canonical nature of the transformation, used with an equal-space Poisson bracket, gives the r-matrix relation that makes the conserved quantities commute.","core_discovery":"For an AKNS integrable system split at a moving boundary x=c(t), the paper defines the defect condition as a Bäcklund transformation evaluated at that boundary. Its central discovery is that the combined system remains integrable: the generating function I(λ)=∫_{-∞}^{c(t)} ũΓ̃ dx + ∫_{c(t)}^{∞} uΓ dx − ln(B11+B12Γ)|_{x=c(t)} is time-independent, so it produces an infinite family of conservation laws; and for NLS, the piecewise transition matrix built from the equal-space monodromies obeys the same classical r-matrix relation, giving Liouville integrability in the Hamiltonian sense. In the KdV case the authors further find that a defect moving at speed $4k^{2}$ admits the continuous peaked solution u=$2k^{2}$ $sech^{2}$(|ξ|+γ), with a peakon for γ>0 and a two-peak/anti-peakon shape for γ<0.","pith_inferences":["If the r-matrix relation holds as asserted, the same moving-defect mechanism should let one construct solutions on the left by applying a Bäcklund transformation to a bulk solution on the right, so moving-defect analogues of finite-gap and multi-soliton solutions should exist; the paper only sketches this route.","The peaked KdV solution is not a weak peakon in the shallow-water sense, as the paper itself stresses; a natural test is whether similar defect-induced peakons appear for NLS or mKdV, and whether they survive when the defect speed differs from the wave speed.","Because the equal-space Poisson bracket treatment is canonical, the classical r-matrix argument may extend to discrete integrable systems with moving defects, such as an integrable discrete NLS or Toda lattice; the paper mentions this possibility but does not prove it.","The paper establishes integrability of the moving-defect NLS in the equal-space bracket; a further check would be whether the infinite set of conserved quantities remains in involution under the standard equal-time bracket as well."],"forward_implications":["The time-dependent defect NLS, KdV, and mKdV systems each admit infinitely many conservation laws generated by I(λ).","For the NLS defect system, Liouville integrability holds in the equal-space Poisson bracket, via a piecewise transition matrix satisfying the r-matrix relation.","The construction extends to multiple moving defects: with n defect locations c1(t)<...<cn(t), the generating function sums bulk contributions plus defect logarithms, and integrability persists.","The KdV equation with a defect at x=4k^2t has the peaked solution 2k^2 sech^2(|k(x-4k^2t)|+γ), with discontinuous first derivative at the peak, and multiple defects produce M-shape peakon/anti-peakon trains.","The defect systems admit Lagrangian descriptions with defect terms involving c'(t); setting c'(t)=0 recovers the fixed-defect results."],"supporting_citations":[{"why":"introduces the systematic defect-as-BT construction and the generating function for conserved quantities that Proposition 1 generalizes to a moving boundary.","marker":"[13]"},{"why":"supplies the equal-space Poisson bracket and the r-matrix relation for a fixed defect, which Section 4.2 adapts to the time-dependent NLS defect.","marker":"[14]"},{"why":"provides the Bäcklund-transformation defect conditions and Lagrangian defect terms for NLS and the fixed-defect case recovered when c'(t)=0.","marker":"[9]"},{"why":"notes that a defect can move with constant speed, motivating the time-dependent location x=c(t).","marker":"[7]"},{"why":"gives the AKNS spectral problems on which the Lax pairs, conservation laws, and Bäcklund transformations for NLS, KdV, and mKdV are built.","marker":"[19]"},{"why":"introduces peaked solitons for the shallow-water equation that serves as the standard peakon comparison class.","marker":"[25]"},{"why":"establishes the canonical-transformation interpretation of Bäcklund transformations used to transfer the r-matrix argument to the defect system.","marker":"[27]"}],"fun_headline_variants":["Moving Bäcklund defects keep soliton equations integrable","Time-dependent defects preserve integrability, yield peakons","Defect at moving boundary: integrability intact, peakons appear","Soliton integrability survives moving Bäcklund defects","Peakons from moving defects in integrable soliton equations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a Bäcklund transformation evaluated at a moving point is still canonical, so the piecewise transition matrix obeys the same commutativity relation as the defect-free system; the paper asserts this without displaying the boundary computation, and if extra boundary terms appear the integrability claim fails.","fun_headline_variants_meta":{"raw":{"variants":["Moving Bäcklund defects keep soliton equations integrable","Time-dependent defects preserve integrability, yield peakons","Defect at moving boundary: integrability intact, peakons appear","Soliton integrability survives moving Bäcklund defects","Peakons from moving defects in integrable soliton equations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000307,"raw_usage":{"total_tokens":1695,"prompt_tokens":823,"completion_tokens":872,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":439,"completion_tokens_details":{"reasoning_tokens":789}},"tokens_in":439,"tokens_out":872,"duration_ms":7472,"temperature":1.0,"reasoning_tokens":789,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:08:51.021455+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the equal-space Poisson bracket {M1(x,t,λ), M2(x,t,μ)} explicitly for the piecewise-defined transition matrix near the moving boundary x=c(t) with c'(t)≠0. If the result differs from [r(λ−μ), M(x,t,λ)⊗M(x,t,μ)] by terms supported at the boundary, the Liouville-integrability claim for the time-dependent defect NLS system collapses. Short of that analytic check, a numerical test of the first few conserved quantities for the defect KdV equation with a non-constant c(t) would provide evidence.","supporting_citations":[{"cited_title":"Caudrelier, On a systematic approach to defects in classical integrable field theories, Int","cited_arxiv_id":null,"evidence_quote":"introduces the systematic defect-as-BT construction and the generating function for conserved quantities that Proposition 1 generalizes to a moving boundary."},{"cited_title":"Caudrelier and A","cited_arxiv_id":null,"evidence_quote":"supplies the equal-space Poisson bracket and the r-matrix relation for a fixed defect, which Section 4.2 adapts to the time-dependent NLS defect."},{"cited_title":"Corrigan and C","cited_arxiv_id":null,"evidence_quote":"provides the Bäcklund-transformation defect conditions and Lagrangian defect terms for NLS and the fixed-defect case recovered when c'(t)=0."},{"cited_title":"Bowcock, E","cited_arxiv_id":null,"evidence_quote":"notes that a defect can move with constant speed, motivating the time-dependent location x=c(t)."},{"cited_title":"Ablowitz, D.J","cited_arxiv_id":null,"evidence_quote":"gives the AKNS spectral problems on which the Lax pairs, conservation laws, and Bäcklund transformations for NLS, KdV, and mKdV are built."},{"cited_title":"Camassa and D","cited_arxiv_id":null,"evidence_quote":"introduces peaked solitons for the shallow-water equation that serves as the standard peakon comparison class."},{"cited_title":"Kodama and M","cited_arxiv_id":null,"evidence_quote":"establishes the canonical-transformation interpretation of Bäcklund transformations used to transfer the r-matrix argument to the defect system."}],"review_version":1}