{"id":"c582911b-aa41-49b3-8179-b4171e381a6e","arxiv_id":"2608.08249","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A recursive moving frame algorithm that normalizes Lie pseudo-group parameters before computing the prolonged action, avoiding recurrence relations and Maurer-Cartan forms.","lead":"This paper presents a new way to compute moving frames for infinite-dimensional symmetry groups, by applying normalizations to the group action early in the calculation. This keeps intermediate symbolic expressions small, making the method more practical for computer algebra.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Recursive normalization lacks a transversality guarantee: full rank of the classical normalization system does not imply that each selected derivative in (4.27) is solvable.","rationale":"The reader's weakest-assumption identification is correct and is the most load-bearing gap in the paper. The recursive algorithm is not backed by a transversality lemma for the restricted equations (4.27)-(4.29); the classical Implicit Function Theorem argument for (3.2) does not transfer automatically to the sequential, one-derivative-at-a-time procedure. This directly threatens the central claim that normalizations can be performed before computing the prolonged action and that the order of normalization does not matter. The six worked examples are internally consistent and reproduce previously known invariants, which is genuine supporting evidence, but they do not establish a general theorem. The foundational convergence and reduced-determining-equation results cited as [40] and [42] are same-author unpublished preprints, so the general correctness claim rests on unverified external foundations. For these reasons a full rejection would be too harsh, while unconditional acceptance would be premature. The reader's CONDITIONAL verdict remains appropriate.","tokens_in":19779,"tokens_out":8593,"duration_ms":94857,"concrete_test":"Implement the recursion symbolically for a family of free, regular Lie pseudo-group actions, for example X=x+a, Y=y+b, U=u+f(x)y+g(x) with suitable normalizations, and at each step compute the linearization coefficient dN^alpha/dZ^a_J at the identity and at the current partial normal form. If the class/order rule ever forces a zero pivot while another normalization order succeeds, the recursion as stated fails. Running the same pivot check on the paper's own Examples 7, 13, and 15 would also reveal whether those successes rely on accidental nonvanishing coefficients rather than a general guarantee.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central recursive step (4.27)-(4.29) requires that the restricted normal form equation N^alpha = 0 can be solved for the chosen parametric derivative Z^a_J after restriction to the hyperplane H_k. The paper proves such solvability only for the simultaneous normalization system (3.2), via the Implicit Function Theorem and transversality of the cross-section. Freeness and regularity of the prolonged action imply that the total Jacobian of all normalization equations with respect to all parametric derivatives has full rank; they do not imply that the particular derivative selected by the class/order rule has a nonzero coefficient in the particular equation used. A full-rank matrix can have a zero pivot in a chosen row/column, in which case the recursive step cannot be solved and the algorithm stalls. The Introduction's claim that 'the order in which these initial conditions are imposed ... does not matter' is never proved; sequential normalizations generally require either a pivoting rule or a commutativity argument. This is load-bearing because it is the only bridge from the classical existence theorem to the recursive procedure.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a recursive implementation of the equivariant moving frame method for infinite-dimensional Lie pseudo-groups. Instead of computing the fully prolonged action before normalizing, the algorithm uses the normal form equation (4.1) and the Rees decomposition of the parametric pseudo-group derivatives into a finite set S and involutive cones C(Z^a_J). Normalizations are performed step by step: the base point is fixed by x=0, then a parametric derivative of largest class and highest order is selected, and the corresponding restricted normal form equation (4.24) or (4.27) is solved for that derivative; prolongation along multiplicative variables normalizes an entire involutive cone, and the result is substituted back before the next step. The authors claim that the order in which normalizations are imposed and the prolonged action is computed does not matter, and that the method avoids recurrence relations and Maurer–Cartan forms. The paper illustrates the algorithm with a running example and with several additional examples, including a non-quasi-horizontal action, a subcone normalization, and a finite-dimensional group action recovering the Schwarzian derivative.","tokens_in":19979,"tokens_out":9614,"duration_ms":86640,"significance":"If the central correctness claims are established, this is a useful algorithmic contribution: it has the potential to make moving frame computations for Lie pseudo-groups substantially more tractable, and the worked examples are convincing evidence that the method works in practice. The paper explicitly reproduces previously known differential invariants, including the Schwarzian derivative and invariants from [38,40], which is a concrete strength. The exposition is clear and the examples are detailed, including the subtle case of normalizing a subcone of an involutive cone. However, the paper does not provide a proof of the two load-bearing claims: solvability of each restricted recursive normalization equation, and independence of the order of normalizations. The correctness of the algorithm is therefore asserted rather than demonstrated. The paper also relies on the authors' own preprints [40,42] for foundational results on reduced determining equations and convergence of normal form power series, which should be stated explicitly.","major_comments":[{"comment":"The recursive step requires that the restricted normal form equation (4.27) can be solved for the selected parametric derivative Z^a_J after restriction to the hyperplane H_k. The paper justifies solvability only for the full simultaneous system (3.2), via the Implicit Function Theorem and transversality of the cross-section. Freeness and regularity of the prolonged action imply full rank of the total Jacobian of all normalization equations with respect to all parametric derivatives; they do not imply that the particular derivative chosen by the class/order rule has a nonzero coefficient in the particular equation used. A full-rank matrix can have a zero pivot in a chosen row/column, in which case the recursive step cannot be solved and the algorithm stalls. The authors need either a recursive transversality lemma, showing that after previously normalized derivatives are substituted the selected derivative is always solvable, or an explicit pivoting rule together with a proof that a valid pivot always exists. This is load-bearing because it is the only bridge from the classical existence theorem to the recursive procedure.","section":"§4, Eqs. (4.27)–(4.29)"},{"comment":"The central claim that 'the order in which these initial conditions are imposed on the pseudo-group transformation and when the prolonged action is computed does not matter' is asserted in the Introduction but is never formalized or proved. Sequential normalizations are generally order-dependent unless a commutativity or confluent termination argument is supplied. The manuscript should state and prove that any admissible choice of parametric derivative in (4.24)–(4.29), or any allowed order of normalizations, yields the same moving frame and the same differential invariants, or at least that the final normal form and the set of unnormalized Taylor coefficients are independent of the choices. Currently the claim rests on the worked examples rather than on a general argument.","section":"Introduction and §4"},{"comment":"Termination and completeness are only asserted: 'Assuming that the prolonged action eventually becomes free, the process of normalizing pseudo-group parameters will eventually terminate at a finite order n with the creation of a moving frame.' No argument is given that the recursive normalization of involutive cones and of the finite set S exhausts Z_par before infinite regress, nor that the residual normal form equations not used for normalization produce a complete set of differential invariants. A theorem with proof is needed, since the examples only demonstrate the procedure for particular actions. The proof should in particular show that each normalization of a cone C(Z^a_J) by prolongation along multiplicative variables is compatible with previously imposed normalizations on overlapping cones, and that no derivative is normalized twice with conflicting constants.","section":"§4, after Eq. (4.30)"},{"comment":"The correctness of the algorithm depends on results quoted from the authors' own preprints [40,42]: the existence and involutivity of the reduced determining equations, the reducibility of the pseudo-group when the prolonged action is free, and the convergence of the normal form power series. The manuscript states these as facts but does not restate them as theorems or provide proofs. If these preprints are not yet published, the present paper's theoretical grounding is incomplete. The authors should either include the precise statements of the results they use, with proofs or pointers to peer-reviewed versions, or clearly mark which results are assumed from the preprints.","section":"§2 and §3"}],"minor_comments":[{"comment":"There is a typo in the first paragraph: 'Building of the works of Cotton' should read 'Building on the works of Cotton'.","section":"Introduction"},{"comment":"The word 'prescibed' should be 'prescribed' in the sentence 'where X0, Y0, Z0, U0 determine the base point on the prescibed hypersurface'.","section":"Example 13"},{"comment":"The phrase 'group of linear factional transformations' should be 'group of linear fractional transformations'.","section":"Example 16"},{"comment":"Figure 1 is a useful diagram, but it is not numbered or captioned in the text; adding a proper caption would improve readability.","section":"§4, Fig. 1"},{"comment":"The algorithm is described in prose rather than as a numbered pseudocode block. A concise pseudocode summary would make the recursive procedure easier to implement and to compare with the classical algorithm.","section":"§4"},{"comment":"References [40] and [42] are listed as preprints; if they have been updated or accepted, the published versions should be cited.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper's central gap is the missing proof of recursive solvability and order independence; this is fixable but requires substantial theorem-level work. The examples are strong and suggest the method is correct. The reliance on two same-author preprints [40,42] for foundational results is worth verifying with the editor; if those preprints are not yet accepted, the theoretical underpinning of the present paper is incomplete. The paper fits the scope of math-ph reasonably well, though it is closer to differential geometry and symbolic computation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the recursive moving frame algorithm built on normal form equations and Rees decompositions, avoiding the recurrence relations and Maurer-Cartan forms used in the authors' earlier recursive schemes. That is not just a cosmetic variant: it lets you normalize pseudo-group parameters at low order before computing the prolonged action, which is what keeps the symbolic expressions manageable. The six worked examples are clean, internally consistent, and reproduce known invariants, including the Schwarzian derivative. That is real evidence the method works as demonstrated.\n\nThe soft spot is exactly where the stress-test note lands. The recursive step (4.27)-(4.29) requires that the restricted normal form equation can be solved for the particular parametric derivative chosen by the class/order rule. Full rank of the simultaneous normalization system does not guarantee that. You can have a full-rank Jacobian with a zero pivot in the chosen row and column, in which case the recursive step stalls. The paper asserts that 'the order in which these initial conditions are imposed ... does not matter' (Introduction) but never proves it. Sequential normalizations generally need either a pivoting rule or a commutativity argument. This is load-bearing, because it is the only bridge from the classical existence theorem (Section 3) to the recursive procedure.\n\nA secondary issue: the foundational results on reduced determining equations and convergence of normal form series come from same-author preprints [40,42]. That is not disqualifying — the examples are benchmarked against independent known results — but it means the general correctness claim rests on unpublished work.\n\nFor a reader working on symbolic moving frame computations, this paper is useful and worth engaging. The examples alone justify a serious look. I would send it to peer review, with the request that the authors either prove the solvability/termination property under precise assumptions, or explicitly restrict the algorithm's guarantees to cases where a pivot condition holds. As is, the method is plausible and well-illustrated, but the central algorithmic claim is not yet fully supported.","headline":"New recursive moving frame algorithm with real promise, but the central order-independence and solvability claims need proof or precise assumptions.","tokens_in":20492,"tokens_out":1738,"would_cite":true,"duration_ms":17909,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["22F05","53A55"],"pacs":[],"model":"deepseek-v4-flash","headline":"A moving frame for a Lie pseudo-group can be constructed recursively by normalizing parametric derivatives one involutive cone at a time, before the full prolonged action is computed, and the order of the steps does not matter.","keywords":["Differential invariants","involutive cone","Lie pseudo-group","moving frame","recursive implementation","normal form equation","Rees decomposition","reduced determining equations"],"falsifier":"A concrete check: impose the same normalizations on the running example (2.2) in the opposite order from Example 7, normalizing $X_x$ before $Y_x$; if the computed differential invariants or the moving frame differ, the order-independence claim is false.","tokens_in":19517,"feed_emoji":"📐","tokens_out":11392,"duration_ms":97542,"temperature":0.7,"pith_summary":"This paper claims that the equivariant moving frame method for infinite-dimensional Lie pseudo-groups can be implemented recursively: normalize the pseudo-group's parametric derivatives one involutive cone at a time, using the normal form equation as the guide, and only then compute the prolonged action on submanifold jets. The central assertion is that the order in which these normalizations are imposed, and the point at which the prolonged action is computed, does not affect the final moving frame or the differential invariants it produces. A sympathetic reader would care because the classical method first prolongs the action to high order, generating expressions that overwhelm symbolic software, while the recursive scheme keeps intermediate expressions manageable. If the claim is right, a complete set of differential invariants can be read off from the normal form's leftover Taylor coefficients without using recurrence relations or Maurer–Cartan forms.","feed_headline":"Recursive method builds moving frames without full prolongation","feed_subtitle":"Normalizing one derivative cone at a time yields differential invariants and keeps expressions small.","key_machinery":"The load-bearing objects are the reduced Lie pseudo-group $\\underline{\\mathcal{G}}$, whose transformations are restricted to the submanifold $s$, its reduced determining equations, and the Rees decomposition $Z^{(\\infty)}_{\\mathrm{par}} = \\mathcal{S} \\cup \\bigcup_{Z^a_J \\in \\mathcal{B}} \\mathcal{C}(Z^a_J)$, which splits parametric derivatives into a finite set plus disjoint involutive cones. An involutive cone $\\mathcal{C}(Z^a_J)$ is the set of all derivatives obtained by repeatedly differentiating $Z^a_J$ with respect to its multiplicative variables $x_1,\\dots,x_k$, where $k$ is the class of the multi-index $J$ under the Pommaret division. The algorithm normalizes one such derivative at a time by solving either the algebraic equations (4.24)–(4.25) or the restricted PDE (4.27)–(4.29) on the hyperplane $H_k$, substitutes the normalization back into the normal form equation, and then prolongs only along non-multiplicative variables. This normalize, substitute, prolong interleaving is what keeps intermediate expressions small and is the reason the paper gives for why the order of the steps does not matter.","core_discovery":"On its own terms, the paper's discovery is that the normal form equation $U(x,u(x)) = \\hat U(X(x,u(x)))$ can be used as a recursive normalization engine. Writing $N^\\alpha = U^\\alpha - \\hat U^\\alpha(X) = 0$, the algorithm repeatedly selects a parametric pseudo-group derivative $Z^a_J$ of largest class and order and solves the corresponding normalization equation: as an algebraic equation at the origin when $Z^a_J$ lies in the finite set $\\mathcal{S}$, or as a PDE restricted to the hyperplane $H_k$ of multiplicative variables when $Z^a_J$ lies in an involutive cone. The result is substituted back into the normal form equation and the whole involutive cone $\\mathcal{C}(Z^a_J)$ (or a subcone) is removed from the Rees decomposition. Because the reduced determining equations are involutive, the normalized values are well-posed initial conditions, and prolongation along the multiplicative variables fills in the entire cone. The paper asserts that the process terminates once the prolonged action becomes free, that the final moving frame and invariant set are independent of the order in which the cones are chosen, and that the leftover normal form Taylor coefficients are a complete set of differential invariants, all without recurrence relations or Maurer–Cartan forms.","pith_inferences":["An implication the authors do not draw is that the order-independence claim could be phrased as a confluence property of the rewriting system defined by the normal form equations; if true, any choice of normalization order would lead to the same normal form.","A testable extension would be to replace the symbolic solution of each restricted equation by numerical integration, producing approximate moving frames and invariants for submanifolds without ever forming the prolonged action.","Because each cone is normalized and then discarded, a practical corollary is that changing the cross-section at low order would only re-run the affected cone normalizations rather than the whole computation.","The paper gives no complexity estimate; a natural next step would be to bound the number of normalization steps in terms of the classes and orders of the parametric derivatives in the Rees decomposition."],"forward_implications":["Symbolic moving frame computations for Lie pseudo-groups can be pushed to higher orders, because the expressions no longer contain the full prolonged action at every intermediate step.","A complete set of differential invariants emerges directly from the non-normalized Taylor coefficients of the normal form, without computing recurrence relations or Maurer–Cartan forms.","The recursion works for actions that are not quasi-horizontal, such as $X=x+a$, $Y=y+b$, $U=f(u)$, with no prior hodograph transformation.","For finite-dimensional Lie group actions the scheme reproduces classical invariants; the linear fractional example yields the Schwarzian derivative as a normal form by-product."],"supporting_citations":[{"why":"Introduces the reduced Lie pseudo-group and the reduced determining equations whose involutivity and well-posed initial conditions the recursive scheme exploits.","marker":"[40]"},{"why":"The standard equivariant moving frame construction for Lie pseudo-groups that the recursion generalizes and the baseline for comparison.","marker":"[38]"},{"why":"Supplies the Pommaret division, Rees decomposition, and involutive cone structure used to organize parametric derivatives.","marker":"[46]"},{"why":"The earlier recursive moving frame algorithm for Lie pseudo-groups, based on recurrence and Maurer–Cartan forms, which this paper's algorithm avoids.","marker":"[41]"},{"why":"Introduces the recursive moving frame idea, adapted here to the pseudo-group setting.","marker":"[34]"},{"why":"Provides the Maurer–Cartan form machinery used by other recursive methods and deliberately not used here.","marker":"[37]"},{"why":"Companion convergence result for normal form power series in the general case, supporting the normal-form-equation viewpoint.","marker":"[42]"}],"fun_headline_variants":["Recursive normalizations build moving frames cone by cone","No full prolongation: recursive moving frames for Lie pseudo-groups","Moving frames via recursive normalizations, skipping full prolongation","Normalize before prolong: recursive moving frames algorithm"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The method assumes that, at every recursive step, the chosen derivative of the transformation can be isolated from the restricted normal form equation; the paper proves this solvability only for the classical one-step construction, not for the recursive equations.","fun_headline_variants_meta":{"raw":{"variants":["Recursive normalizations build moving frames cone by cone","No full prolongation: recursive moving frames for Lie pseudo-groups","Moving frames via recursive normalizations, skipping full prolongation","Normalize before prolong: recursive moving frames algorithm"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00029,"raw_usage":{"total_tokens":1643,"prompt_tokens":835,"completion_tokens":808,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":451,"completion_tokens_details":{"reasoning_tokens":742}},"tokens_in":451,"tokens_out":808,"duration_ms":7914,"temperature":1.0,"reasoning_tokens":742,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T00:13:58.578260+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: impose the same normalizations on the running example (2.2) in the opposite order from Example 7, normalizing $X_x$ before $Y_x$; if the computed differential invariants or the moving frame differ, the order-independence claim is false.","supporting_citations":[{"cited_title":"https://doi.org/10.1007/s00025-018-0818-5","cited_arxiv_id":null,"evidence_quote":"The earlier recursive moving frame algorithm for Lie pseudo-groups, based on recurrence and Maurer–Cartan forms, which this paper's algorithm avoids."},{"cited_title":"https://doi.org/10.1007/s00025-011-0153-6","cited_arxiv_id":null,"evidence_quote":"Introduces the recursive moving frame idea, adapted here to the pseudo-group setting."},{"cited_title":"https://doi.org/10.1007/s00029-005-0008-7","cited_arxiv_id":null,"evidence_quote":"Provides the Maurer–Cartan form machinery used by other recursive methods and deliberately not used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Companion convergence result for normal form power series in the general case, supporting the normal-form-equation viewpoint."}],"review_version":1}