{"id":"9e2f3630-36a5-4fa1-ae6b-4d28e175a87b","arxiv_id":"2607.18312","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A vakonomic variational treatment of Koopman-discretized incompressible Euler equations yields Lie-Poisson Lax dynamics with exact circulation and Casimir preservation in the reset-free limit.","lead":"The paper builds a new computer simulation method for ideal fluids by treating the numerical error from discretization as a nonholonomic constraint and applying a vakonomic variational principle, yielding equations that preserve circulation and Casimir invariants to machine precision. The method runs stably at low resolutions but requires periodic 'resets' that break exact conservation, and its convergence to the true Euler equations is not proven.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim is conditional on the unproven continuum limit: the paper's own Section 12 leaves open whether so(F) and im(Ā) converge to sdiff(M); resetting also breaks the headline invariant conservation in the implemented algorithm.","rationale":"The reader's weakest assumption—that the discrete sub-Riemannian geodesic flow converges to the Euler equations in the continuum limit—is exactly the load-bearing condition for the central claim, and the manuscript itself identifies it as open. I also agree that the resetting procedure is a secondary fragility: it interrupts the exact conservation that motivates the construction, and the reset interval is hyperparameter-tuned rather than derived. The algebraic derivation of the Lax equation and its reset-free conservation laws appears internally sound, and the paper deserves credit for reporting the reset-induced Casimir jumps honestly. However, without a convergence proof or a non-steady convergence study, the method is best regarded as a structure-preserving toy model with promising numerics, not a proven discretization of Euler. This matches the reader's CONDITIONAL verdict, so no verdict change is needed.","tokens_in":56301,"tokens_out":6945,"duration_ms":83507,"concrete_test":"Run a nontrivial unsteady 2D benchmark—e.g., the double shear layer from Sec. 11.4.1—with the IGA–FEEC discretization at resolutions 64^2, 128^2, 256^2, and 512^2, with Δt scaled by CFL, and compare the vorticity field against a well-resolved 1024^2 pseudospectral reference at a fixed physical time. Report the L2 error and the order of convergence. If the error does not decrease systematically with resolution, the claimed consistency with Euler is not supported. Additionally, record the magnitude of Casimir jumps at the first reset to quantify the gap between reset-free theory and the implemented algorithm.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim—that the vakonomic Lax system is a discretization of the incompressible Euler equations—requires the finite-dimensional data (so(F), im(Ā), K) to converge, in an appropriate sense, to (sdiff(M), sdiff(M), L2). The authors explicitly leave this open in Section 12: \"it is interesting to consider whether the Lie algebra so(F) underlying the vakonomic discretization actually converges to sdiff(M) in the continuum limit, and if so at what rate.\" Without such a consistency theorem, Eq. (6.12) is a self-consistent isospectral flow on so(F), but it is not established as an approximation of Eq. (3.9). The Taylor–Green convergence test in Fig. 12 is evidence for one steady solution, not for the general initial-value problem. Moreover, the practical algorithm in Alg. 6 includes a resetting step that discontinuously changes the Clebsch labels (Sec. 10.3); the abstract's \"machine-precision satisfaction of Casimir invariants\" is therefore true only in the reset-free limit, while all production experiments use resets with periods selected by heuristics (Table 2). These two gaps are acknowledged in the manuscript, but they are precisely the conditions needed for the central claim to hold as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a finite-dimensional discretization of the incompressible Euler equations based on a Koopman representation of volume-preserving diffeomorphisms on half-densities. The discretized velocity constraint X ∈ im(Ā) is nonholonomic, and the authors treat it with the vakonomic variational principle, obtaining a reduced Lie–Poisson system in Lax form, Eq. (6.12): Ż = [Z^#, Z]. They show this system is isospectral, hence preserves Casimir invariants, and satisfies a discrete analogue of Kelvin's circulation theorem. A low-rank Clebsch momentum-map representation reduces the computational cost from O(F^2) to O(mF), and a Lie-trapezoidal time integrator preserves energy and the coadjoint orbit in the reset-free case. The practical algorithm includes a periodic resetting procedure to control dispersion. Numerical experiments include Taylor–Green vortex convergence, invariant checks, shielded and leapfrogging vortices, flows on surfaces, a 3D trefoil knot, and a Boussinesq Rayleigh–Taylor simulation.","tokens_in":56713,"tokens_out":5390,"duration_ms":55208,"significance":"The core geometric derivation is clean and largely self-contained: Theorem 6.1 derives the coadjoint equation from the vakonomic action, Lemma 6.1 gives the coadjoint action on so(F), and Theorem 8.1 shows the Clebsch ansatz satisfies the Lax equation exactly. The integrator analysis in Theorem 9.1 is rigorous and the numerical experiments are extensive. If the discrete system were shown to converge to the Euler equations in the continuum limit, this would be an important contribution to structure-preserving fluid simulation. Even without such a convergence theorem, the Lax formulation, the low-rank momentum-map representation, and the numerical evidence of reduced dispersion are valuable. The paper is also commendably explicit about its open questions, especially the continuum-limit issue in Section 12.","major_comments":[{"comment":"The paper's central claim—that the vakonomic Lax system (6.12) is a discretization of the incompressible Euler equations (3.9)—is not established. Section 12 explicitly leaves open whether so(F) and im(Ā) converge to sdiff(M), and the numerical evidence in Fig. 12 is only for the steady Taylor–Green vortex. Without a consistency or convergence result for the general initial-value problem, Eq. (6.12) is a self-consistent isospectral flow on so(F) but not proven to approximate Eq. (3.9). I recommend either supplying such a result or reframing the contribution as a structure-preserving discrete fluid model whose relation to the Euler equations is empirical.","section":"Section 12; Sections 5–6"},{"comment":"The abstract's 'machine-precision satisfaction of Casimir invariants' is true only in the reset-free limit. Algorithm 6 includes resets, all production experiments use resets (Table 2), and Section 11.1.1 reports a 1.6% energy drop over 5 s for TGV with resets; Fig. 16 shows Casimir jumps at each reset. Since resetting is a non-Hamiltonian heuristic (Sec. 10.3.2), the headline conservation claim should be qualified to the reset-free integrator, and the practical algorithm should be presented as approximately structure-preserving with quantified drift.","section":"Abstract; Section 10.3; Section 11.1.1"},{"comment":"The low-rank Clebsch ansatz (8.2) is exact on the invariant manifold rank(Z) ≤ 2m, but the paper does not quantify the approximation error for general initial data that are not low-rank, nor does it prove that the discrete analogue of the d−1 Clebsch-pair sufficiency of [83] holds in the finite-dimensional setting. The paper itself notes 'some variance in practice' (Sec. 12). This is not fatal, but it should be acknowledged explicitly in the main text rather than presenting the low-rank model as a closed reduced-order description with no approximation gap.","section":"Section 8; Section 12"}],"minor_comments":[{"comment":"Heading typo: 'Eqations' should be 'Equations'.","section":"Section 6"},{"comment":"The symbol V is used both for the finite element space and for its dimension. Use, e.g., dim V to avoid ambiguity.","section":"Section 5.4"},{"comment":"The reset-frequency parameter α in Fig. 15 is not defined in the table or caption; clarify whether it is a step count or a time interval.","section":"Figure 15 / Table 2"},{"comment":"The FTLE threshold S_f ≤ 1 is stated to 'consistently lead to good results' but no sensitivity analysis is provided; this is an empirical heuristic and should be labeled as such.","section":"Section 10.3.1"},{"comment":"The comparison baseline 'Vorticity FEEC' is reconstructed via Reset for Casimir diagnosis. Please explain why this reconstruction is faithful to the LdA dynamics; otherwise the comparison in Fig. 14 is difficult to interpret.","section":"Section 11.1.2"},{"comment":"The statement that LdA is 'not self-consistent' is normative; consider rewording to 'not variational in the same intrinsic sense' to avoid overstatement.","section":"Section 7.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong candidate after revision. The variational derivation and integrator analysis are sound, and the numerical work is substantial. The main risks are overclaiming the connection to the Euler equations without a continuum-limit result, and overstating conservation in the presence of resets. If the authors are willing to reframe the contribution accordingly and quantify the reset-induced drift, the paper could become acceptable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline: this is a genuinely new variational route to discrete fluid dynamics, and the theory is cleaner than the abstract's marketing suggests—but the \"machine-precision\" conservation claim should be read as a reset-free statement, and whether this actually approximates incompressible Euler is still open.\n\nWhat's new: the vakonomic treatment of the nonholonomic constraint that arises from Koopman discretization is a real departure from the standard Lagrange–d'Alembert approach. The derivation of the Lax equation (6.12) from the sub-Riemannian geodesic principle is rigorous and self-contained. The discrete Kelvin circulation theorem (Theorem 6.2) is stronger than the previous LdA result, and the low-rank Clebsch parameterization (Theorem 8.1) is a neat practical idea that preserves the Lax structure while cutting the cost from O(F^2) to O(mF). The numerical experiments, particularly the trefoil and leapfrogging cases, show qualitative improvements over baselines, and the paper is honest about its limitations—it explicitly flags the missing continuum limit and the resetting issue.\n\nWhere I'd push back: the missing continuum limit is not a minor technicality. The paper's central claim is that this is a discretization of the Euler equations, but Section 12 leaves open whether so(F) and im(A-bar) converge to sdiff(M). Without that, Eq. (6.12) is a self-consistent isospectral flow on a finite-dimensional space, but it's not established that it approximates Eq. (3.9). The Taylor–Green convergence test is evidence for one steady solution, not for the general initial-value problem. That said, this is not a fatal flaw for all readers—many structure-preserving schemes are used before convergence is proved—but it needs to be stated as an open question rather than an incidental remark.\n\nThe resetting issue is more concrete. The headline conservation of Casimirs holds only between resets; the actual algorithm resets frequently (every 0.21 s in most experiments), and each reset causes a jump. The energy drops about 1.6% over 5 s in one test, and the reset period is chosen per experiment by heuristics. The paper acknowledges this, but it means the abstract's \"machine-precision satisfaction\" is not what the production code delivers. That's a mismatch between the claim and the implementation.\n\nThe lack of released code is a minor concern for verifiability, but the paper describes the algorithm in enough detail that reimplementation is feasible.\n\nVerdict: this is a serious paper with a real idea, and it deserves a serious referee. If I were the editor, I'd send it to peer review but ask the referees to weigh the continuum limit question and the resetting issue carefully. The authors should either prove convergence under reasonable assumptions or soften the claim to \"a structure-preserving discrete model inspired by Euler,\" and they should present the reset-free conservation as the theoretical property, with resetting as a practical add-on.","headline":"Clean vakonomic Lax derivation and a strong discrete Kelvin theorem, but the abstract oversells conservation and the continuum limit is unproven.","tokens_in":57138,"tokens_out":2409,"would_cite":true,"duration_ms":31186,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["70F25","76B03","37K10","53D20","65M60"],"pacs":[],"model":"deepseek-v4-flash","headline":"Forcing every variation to respect the discretized velocity constraint yields fluid paths that are true geodesics, conserving energy, enstrophy-like invariants, and circulation to machine precision.","keywords":["vakonomic mechanics","incompressible Euler equations","Koopman representation","sub-Riemannian geodesics","Lax equation","Lie–Poisson systems","Clebsch variables","Kelvin's circulation theorem"],"falsifier":"Run the reset-free scheme on a Taylor–Green vortex at 32→64→128→256 resolution and measure the L^2 error of the vorticity field against the exact steady solution; if the error does not shrink at a stable order, the dynamics are structure-preserving but not a discretization of Euler. Alternatively, run a two-dipole leapfrogging experiment with resetting disabled: if tr(Z^2) and tr(Z^4) hold to machine precision while the vortex trajectories visibly miss the point-vortex prediction, then the exactly conserved structure is not the one that carries the fluid physics.","tokens_in":56215,"feed_emoji":"🌊","tokens_out":11343,"duration_ms":97448,"temperature":0.7,"pith_summary":"Incompressible Euler flow is classically the geodesic equation on the group of volume-preserving diffeomorphisms; this paper asks what survives of that geodesic picture when the flow is discretized through the Koopman representation of diffeomorphisms as rotation matrices. That discretization imposes a nonholonomic constraint on admissible velocities, and the paper shows that enforcing it the vakonomic way — allowing only constraint-respecting variations — keeps the discrete flow a genuine geodesic on a sub-Riemannian matrix manifold. The stationarity condition closes as a Lax equation, dZ/dt = [Z^#, Z], which makes the system Lie–Poisson: kinetic energy, the entire spectrum (the Casimir invariants, discrete analogues of enstrophy and helicity), and a discrete Kelvin circulation theorem are all conserved to solver tolerance. A low-rank Clebsch momentum-map representation reduces the per-step cost from O(F^2) to O(mF) without sacrificing any of the structure, and experiments show stable, low-dispersion vortex evolution on coarse grids, on surfaces, and in 3D. The claim that would make this a full discretization theory — that the discrete dynamics converges to the Euler equations as the mesh refines — is stated and left open.","feed_headline":"Discrete fluid flows conserve circulation to machine precision","feed_subtitle":"Constrained-geodesic variation yields a Lax flow preserving energy and enstrophy-like invariants on coarse grids.","key_machinery":"The load-bearing object is the vakonomic variational principle applied to the right-invariant K-metric restricted to the distribution D_R = {XR : X ∈ im(A-bar)} in T SO(F). Lagrange–d'Alembert lets variations leave the distribution and therefore requires an ad hoc ambient-metric choice; the vakonomic principle instead confines the whole variation family to admissible paths, and the stationarity condition then closes as the Lax equation dZ/dt = [Z^#, Z]. The sharp operator # = A-bar K^-1 A-bar* is defined solely from the discrete advection operator and the velocity mass matrix K, so no metric information from outside the constraint space is needed — the heart of the paper's self-consistency c","core_discovery":"Set up discrete fluid motion as a path R(t) in SO(F) whose body velocity X = dR/dt R^-1 is constrained to the image im(A-bar) of a discrete advection operator; because im(A-bar) is not closed under the matrix commutator, this is a genuine nonholonomic constraint. The paper proves that stationarity of the kinetic-energy action under variations that keep the family inside the constraint distribution — the vakonomic principle, as opposed to Lagrange–d'Alembert — is equivalent to a coadjoint evolution on so(F)* that reads, after the Frobenius identification, as the matrix Lax equation dZ/dt = [Z^#, Z], with sharp map # = A-bar K^-1 A-bar* built only from constraint-space data. This evolution is","pith_inferences":["My inference: the reset step, which the paper needs to curb dispersion, breaks the exact conservation the construction advertises; a natural testable extension is the paper's own suggestion of double-bracket dissipation, which would limit dispersion while keeping Casimirs exactly conserved.","My inference: the advertised clean separation of time-integration drift from resetting drift is only implicit in the figures; a reset-free diagnostic run that logs tr(Z^2) and tr(Z^4) every step would settle how much of the observed dissipation is actually the reset's doing.","My inference: the circulation theorem's physical content depends on how faithfully the Lie-algebraic closure of im(A-bar) represents real material loops; a coarse-mesh reset-free run against analytical point-vortex or shielded-vortex trajectories would quantify that fidelity.","My inference: the m = d choice of Clebsch pairs sets a fixed rank for Z, and the paper gives no rank-robustness test; sweeping m while monitoring the reconstructed velocity would show whether truncation in the low-rank ansatz is a source of error comparable to the discretization error."],"forward_implications":["Discrete ideal-fluid simulations become exactly isospectral: enstrophy- and helicity-like Casimirs no longer drift, removing a common source of spurious energy cascades in long runs.","One source of arbitrariness vanishes: unlike Lagrange–d'Alembert discretizations, the equations carry no free 'ambient metric' — only the choice of function space, velocity space, and mass matrix.","The low-rank Clebsch form reduces the per-step cost from O(F^2) to O(mF), making exactly structure-preserving ideal-fluid simulation practical on standard FEEC/B-spline and triangle-mesh grids without ever assembling a dense matrix.","The conserved discrete circulation is stronger than the weak circulation of earlier projected-Lax schemes: it holds for every loop transported by the flow, not merely in a projected sense.","The same variational framework extends to semidirect-product physics — stratified Boussinesq flow is demonstrated, and the authors sketch shallow-water, compressible, and MHD variants."],"fun_headline_variants":["Vakonomic variation preserves circulation on coarse grids","Discrete fluid geodesics yield machine-precision circulation","Nonholonomic fluid dynamics with exact invariants","Geodesic discretization conserves Casimirs on coarse grids","Vakonomic fluid flow: exact circulation without fine grids"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The whole construction rests on the premise that the constrained rotation-matrix dynamics actually approaches the incompressible Euler equations as the grid is refined — the paper proves exact structure preservation but explicitly leaves the convergence of its discrete Lie algebra to the divergence-free vector fields open; if that convergence fails, the method is a self-consistent toy rather than a discretization of Euler flow.","fun_headline_variants_meta":{"raw":{"variants":["Vakonomic variation preserves circulation on coarse grids","Discrete fluid geodesics yield machine-precision circulation","Nonholonomic fluid dynamics with exact invariants","Geodesic discretization conserves Casimirs on coarse grids","Vakonomic fluid flow: exact circulation without fine grids"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000428,"raw_usage":{"total_tokens":2016,"prompt_tokens":723,"completion_tokens":1293,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":467,"completion_tokens_details":{"reasoning_tokens":1229}},"tokens_in":467,"tokens_out":1293,"duration_ms":9926,"temperature":1.0,"reasoning_tokens":1229,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T01:35:25.340731+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the reset-free scheme on a Taylor–Green vortex at 32→64→128→256 resolution and measure the L^2 error of the vorticity field against the exact steady solution; if the error does not shrink at a stable order, the dynamics are structure-preserving but not a discretization of Euler. Alternatively, run a two-dipole leapfrogging experiment with resetting disabled: if tr(Z^2) and tr(Z^4) hold to machine precision while the vortex trajectories visibly miss the point-vortex prediction, then the exactly conserved structure is not the one that carries the fluid physics.","supporting_citations":[],"review_version":2}