{"id":"2b85d39b-79d0-471a-8509-142c7a8c1c60","arxiv_id":"2608.04191","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"An artificial-compressibility relaxation lets PINNs solve smooth incompressible flows with sub-0.5% velocity error, while vortex shedding is only recovered when sparse data from a finite-element reference is assimilated.","lead":"This paper tests a physics-informed neural network that replaces the hard zero-divergence rule in fluid flow with a soft relaxation controlled by a single number, epsilon. On smooth test flows the method is accurate and stable, but on the classic cylinder wake it fails unless sparse sensor data from a separate simulation is assimilated.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Cylinder wake validation is circular: sensors and ground truth derive from the same FEM reference with Strouhal 0.176 (above the 0.164–0.172 band), so the 7%/3% accuracy claims do not independently establish fidelity to the true flow.","rationale":"The paper's central positive claim for unsteady flow is the data-assimilation recovery of the cylinder wake, and that claim rests on validation against the same FEM simulation that supplied the sensors (Sections 7.1–7.2). The reader's weakest_assumption already identifies the FEM reference Strouhal of 0.176 as a load-bearing target, and the reader's rationale explicitly flags the circularity. My stress-test agrees that this is the most consequential weakness: the forward collapse in Section 7 shows that low residual alone does not select the unsteady branch, so the sensors are doing the branch-selection work, and the subsequent error metrics are measured against the reference that generated those sensors. The authors are commendably transparent about this limitation, which is why the verdict remains CONDITIONAL rather than REJECT, but the abstract's 'within 3%' phrasing can easily be misread as an accuracy claim against the true shedding frequency. A concrete way to settle the concern is to repeat the assimilation with an independent, literature-consistent reference. A secondary observation: the abstract and Section 5.4 state that the fixed-epsilon experiments span 'five decades' of epsilon, but Table 2 lists values from 1e-1 to 1e-4, which is three decades; this numerical overstatement should be corrected in a revision. Since the reader's CONDITIONAL verdict already requires an independent reference (among other conditions), my concern does not change the verdict.","tokens_in":9950,"tokens_out":10586,"duration_ms":95822,"concrete_test":"Run the Section 7.2 assimilation using an independent high-fidelity reference, such as a spectral-element or Lattice-Boltzmann solver whose Strouhal lies inside the 0.164–0.172 band, to generate both the sparse sensors and the error metric. If the AC-PINN matches the independent reference to comparable wake-band error and a shedding frequency within a few percent of the literature band, the circularity concern is resolved. If it only matches the original FEM reference, the validation is reference-bound and the accuracy claims should be restated as relative to the generating solver, not to the true flow.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest positive result for unsteady flows is the data-assimilation recovery of the cylinder wake in Section 7.2. The sparse sensors are read from the boundary-layer-resolved FEM reference of Section 7.1, and the reconstruction error (7% wake-band velocity, 3% shedding frequency) is computed against that same FEM field. This is a circular validation: the AC-PINN is compared to the very data source that guided it, so the reported numbers measure how well the network fits the reference, not how faithfully it reproduces the true flow. The circularity is amplified by the reference's own bias: Table 3 shows the FEM Strouhal is 0.176, above the accepted 0.164–0.172 band, and the authors explicitly state that the 3% figure is 'a bound set by the reference's own fidelity rather than an independent validation against the true flow.' A PINN recovering St = 0.181 while the true flow sheds near 0.168 would be off by roughly 7%, even though the paper's headline says 'within 3%' of the reference. Because the forward solve collapses to the steady symmetric branch (Section 7), the method's ability to reach the unsteady branch is demonstrated only relative to the reference that seeded it. The claim that 'a few hundred sparse sensors ... recover the unsteady vortex street' is therefore not independently established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes an artificial-compressibility physics-informed neural network (AC-PINN) for unsteady incompressible Navier–Stokes equations. The divergence-free constraint is replaced by the relaxed continuity equation ε ∂_t p + ∂_x u + ∂_y v = 0 (Eq. 4), so that every governing equation carries a time derivative and all PINN residuals become local. The paper reports three benchmark studies: Taylor–Green vortex, where a seeded study over five decades of ε shows sub-0.5% velocity error and divergence scaling; lid-driven cavity at Re = 100, where a smoothed-lid unsteady solve reaches 7.7% relative-L2 agreement with Ghia et al. centerline profiles after a sufficiently long time window; and cylinder wake at Re = 100, where the plain forward solve collapses to the steady symmetric branch, data assimilation from a finite-element reference recovers the wake to 7% velocity error and Strouhal number to within 3% of that reference, and viscosity inference requires boundary-layer sensors. The paper is explicitly honest about its negative results and about the reference-fidelity limitations of the cylinder validation.","tokens_in":10213,"tokens_out":5312,"duration_ms":51660,"significance":"The paper's main value is its systematic and unusually transparent characterization of a simple formulation modification, including clear negative results. The Taylor–Green experiments are strengthened by three-seed statistics and by the observation that plain fixed-ε training is stable across five decades, which is a useful practical fact for practitioners. The explicit identification of the forward cylinder failure and the sensor-placement dependence of viscosity identification are genuinely informative. The authors also state plainly that the cylinder validation is circular and that the FEM reference itself sits above the literature Strouhal band. However, the central methodological claim—that artificial compressibility makes incompressible-flow PINN training easier—is not supported by any same-budget comparison against a standard PINN or a hard-constrained alternative, and the cavity benchmark is modified by a smoothed lid, so the reported agreements with classical references are not as clean as they appear. These gaps need to be addressed before the paper's conclusions can be fully accepted.","major_comments":[{"comment":"The cavity benchmark is altered by the smoothed lid profile u_lid(x) = tanh(50x) tanh(50(1−x)), so the problem actually solved is not the sharp-lid Ghia benchmark used for comparison. The reported 7.7% centerline error therefore conflates the effect of the smoothing (a free parameter that is not studied) with the method's accuracy. The authors should either solve the sharp-lid problem, possibly with an appropriate corner treatment, or reformulate the claim as agreement with a smoothed-lid reference and provide a sensitivity study of the smoothing width.","section":"Section 6"},{"comment":"The cylinder data-assimilation validation is circular: the sparse sensors are drawn from the same finite-element reference against which the 7% velocity error and 3% Strouhal error are computed, and that reference has St = 0.176, above the accepted 0.164–0.172 literature band. Although the text repeatedly acknowledges this, the abstract's 'within 3%' statement is an agreement with a biased reference, not with the true flow. The unsteady-recovery claim would be substantially strengthened by comparing against an independent reference solution or by explicitly reporting the error relative to the literature Strouhal band.","section":"Sections 7.1–7.2"},{"comment":"No comparison is made against a standard PINN with a divergence penalty, a streamfunction parameterization, or another hard-constrained architecture under the same network size, collocation count, and training budget. Since the paper's motivation is that the algebraic divergence constraint is stiff and that artificial compressibility makes the residuals local, the empirical claims in Section 5.5 ('none of the advanced fixes were strictly necessary') are absolute statements about the AC-PINN and do not establish an advantage over existing formulations. A same-budget baseline comparison is needed to support the central methodological contribution.","section":"Section 5"}],"minor_comments":[{"comment":"The exact number of sensors, their placement distribution, and whether the sensor data are noiseless are not specified; the abstract says 'a few hundred' but the text gives no precise count, which hinders reproducibility.","section":"Section 7.2"},{"comment":"The inverse-problem results are reported as single runs without seed variability or uncertainty estimates; given the paper's otherwise careful use of seeds in the Taylor–Green study, the viscosity-identifiability conclusion would be more robust with multiple seeds.","section":"Table 4 and Section 7.3"},{"comment":"The final row of Table 2 compares the ε-continuation run (16,000 Adam epochs plus L-BFGS) with fixed-ε runs (4,000 Adam epochs plus L-BFGS); the text notes the budget difference, but the factor-of-three improvement should be explicitly described as partly a training-budget effect, not purely an effect of continuation.","section":"Table 2 and Section 5.4"},{"comment":"The Strouhal numbers are reported to three decimal places, but no uncertainty or statistical variability from the finite-element snapshots is given; a short statement on the temporal resolution of the Strouhal estimate would help the reader judge the 3% figure.","section":"Section 7.1, Table 3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is well-written and honest, and the negative results are valuable. The main barriers to acceptance are the missing baseline comparison against standard PINN formulations and the validity of the cavity and cylinder benchmark comparisons. These are fixable in revision, and I would be willing to review a revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nYou should know about this one: it is a methods paper that maps the artificial-compressibility (AC) relaxation for PINNs on the unsteady incompressible Navier–Stokes equations. The new empirical content is a quantitative epsilon-scaling study on Taylor–Green (five decades, multiple seeds), a documented forward-solve collapse to the steady symmetric cylinder wake, and a sensor-placement identifiability study for viscosity inference. Those are genuinely useful and I do not think they appear in the prior literature.\n\nWhat it does well: the Taylor–Green tests are clean. The residual divergence scales as epsilon*|partial_t p|, which is an algebraic consequence of equation (4), but they verify it and show that fixed-epsilon training is stable down to 1e-4 with sub-0.5% velocity error. Seed-averaged numbers are given. The cavity result is honest: 7.7% centerline error, with the physical time horizon identified as the controlling factor, plus a reported negative result that wall clustering makes it worse. The cylinder forward collapse is also honestly reported and consistent with prior findings.\n\nThe soft spots are real but disclosed. The cylinder data-assimilation validation is circular: the sensors and the ground truth come from the same FEM reference, so the 7% velocity error and 3% frequency match measure how well the network fits that reference, not how close to the true flow. The authors say exactly this in the abstract and Section 7.2, so it is not a hidden flaw. Still, the reference Strouhal is 0.176, above the accepted 0.164–0.172 band, so the 3% claim could correspond to a ~7% error against the true shedding frequency if the literature band is right. A converged-mesh Strouhal check inside the band would strengthen the claim. Also, there is no code or data, and no same-budget comparison against a hard-constrained PINN baseline, so the claimed advantage of AC over divergence-penalty or stream-function formulations is not established. That latter point is minor because the paper mostly claims to map the trade-off, not to beat baselines.\n\nOverall: a sincere and useful methods paper. The central scaling story on Taylor–Green holds up, and the honest documentation of failures is valuable. The inverse sensor-placement study is a nice practical lesson. I would send it to peer review and ask for code/data and ideally an independent Strouhal check before accepting. I would bring it to our reading group when we discuss PINN training pathologies.","headline":"A genuinely useful and honest map of the AC-PINN trade-off, with clean Taylor–Green scaling; the cylinder validation is circular but disclosed, and the paper deserves serious review.","tokens_in":10743,"tokens_out":2019,"would_cite":true,"duration_ms":18218,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q30","65M60","68T07","76D05"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that replacing the divergence-free constraint with an artificial-compressibility relaxation makes unsteady incompressible Navier–Stokes equations trainable for PINNs, and that the method's reach is bounded by…","keywords":["artificial compressibility","physics-informed neural networks","incompressible Navier-Stokes equations","Taylor-Green vortex","lid-driven cavity","cylinder wake","vortex shedding","parameter identification"],"falsifier":"Train the plain forward AC-PINN on the Re=100 cylinder with no data-misfit term and a strong, sustained symmetry-breaking inflow, then measure the wake probe's transverse velocity RMS and spectral peak; the paper predicts collapse to the steady branch with RMS around 3e-3 and no Strouhal peak near 0.17. If the network sheds with Strouhal approximately 0.17 without data, the claim that data-free forward solves cannot reach the instability is falsified.","tokens_in":9718,"feed_emoji":"🌊","tokens_out":7360,"duration_ms":60457,"temperature":0.7,"pith_summary":"Artificial compressibility, the replacement of the divergence-free condition by $\\varepsilon\\,\\partial_t p + \\nabla\\cdot u = 0$, converts the stiff differential-algebraic constraint of incompressible flow into a purely local evolution residual that a physics-informed neural network can minimize directly. The paper establishes this on three benchmarks of increasing difficulty. On the Taylor–Green vortex, the residual divergence scales as $\\varepsilon|\\partial_t p|$, and fixed-$\\varepsilon$ training is stable across five decades of $\\varepsilon$ with sub-0.5% velocity error, so no elaborate training aids are required on smooth problems. On the lid-driven cavity, a long enough time window recovers reference centerline profiles to 7.7%. The boundary of the method is the cylinder wake: the forward solve collapses to the steady symmetric branch, while assimilating a few hundred sparse velocity sensors recovers the vortex street to 7% and its frequency to within 3% of the reference, a bound set by the reference solver's own fidelity. The paper's larger point is that this relaxation maps where a lightweight toolbox suffices and where data assimilation is unavoidable.","feed_headline":"Artificial compressibility makes incompressible flow PINN-trainable","feed_subtitle":"Sub-0.5% errors on smooth flow; unsteady wakes only after sparse velocity data are added.","key_machinery":"The load-bearing object is the relaxed continuity equation $\\varepsilon\\,\\partial_t p + \\partial_x u + \\partial_y v = 0$, with $\\varepsilon$ playing the role of an inverse squared artificial sound speed. It removes the differential-algebraic constraint and makes every governing equation carry a time derivative, so the residual system becomes purely local; the divergence error then scales as $\\varepsilon|\\partial_t p|$. This single parameter carries the argument by trading modelling error against optimisation stiffness, and the paper maps that trade-off on a closed-form benchmark and then applies it to flows where the steady branch competes with an unsteady one.","core_discovery":"The central claim is that the artificial-compressibility relaxation with a single scalar $\\varepsilon$ turns the global divergence constraint into a local residual, at a modelling-error price controlled by $\\varepsilon$ and quantified by the scaling $\\partial_x u + \\partial_y v = -\\varepsilon\\,\\partial_t p$. The paper shows that both divergence and velocity error decrease monotonically and saturate as $\\varepsilon\\to 0$, with no stiffness-induced degradation down to $\\varepsilon=10^{-4}$ on a smooth benchmark. On the $Re=100$ cylinder wake the data-free forward AC-PINN relaxes to the steady symmetric wake, a valid low-residual solution; this failure is not a training artifact but a structural property, so sparse-sensor data assimilation is required to recover shedding. The paper also claims that assimilation fidelity is bounded by the reference fidelity, and that viscosity is identifiable only when sensors are placed in the boundary layer where the viscous term is significant.","pith_inferences":["The same relaxation may transfer to other DAE-constrained PDEs where a global constraint couples all collocation points, giving a local residual at the cost of a tunable modelling error; this is testable by comparing convergence on a problem with a known constraint.","The collapse to the steady branch suggests a testable fix for forward instability: adding a term that penalizes low-residual steady solutions, or explicitly breaking symmetry in the loss landscape, may let a data-free solve reach shedding without sensors.","The sensor-placement result implies an active-learning strategy for inverse problems: place observations where the term to be inferred is largest, such as the boundary layer for viscosity, and use a prior or denser near-wall data to close residual bias.","The paper's negative result on boundary-layer-adaptive sampling for the cavity suggests that local residual balancing via gradient-norm weighting can be destabilized by adding near-wall terms, pointing toward curriculum or hard-constraint treatments of thin layers."],"forward_implications":["On smooth problems with known boundary and initial data, the AC-PINN needs only variable scaling and a plain fixed-$\\varepsilon$ residual to reach sub-0.5% velocity error; adaptive weighting and $\\varepsilon$-continuation improve accuracy but are not necessary.","For steady-limit flows in the lid-driven cavity, the physical time horizon, not collocation density or near-wall sampling, determines accuracy; wall-clustered sampling actively degrades the result.","A data-free forward PINN cannot be expected to produce self-sustained instabilities such as vortex shedding, because the steady branch is a low-residual solution; sparse velocity observations are sufficient to recover the unsteady branch.","Any validation of a data-driven wake reconstruction is only as good as the reference solver's boundary-layer resolution; a reference biased high in Strouhal passes that bias on to the learned solution.","Inverse parameter recovery depends on sensor placement, not sensor count: viscosity estimation fails from wake sensors alone and becomes identifiable once boundary-layer sensors are included."],"supporting_citations":[{"why":"introduces the artificial-compressibility idea from which the relaxed continuity equation is taken.","marker":"[1]"},{"why":"establishes the physics-informed neural network framework used for all residual minimization.","marker":"[3]"},{"why":"documents gradient pathologies from stiff constraints, motivating the local-residual reformulation.","marker":"[6]"},{"why":"reports the same collapse to a steady symmetric branch for a constrained PINN on the cylinder wake.","marker":"[20]"},{"why":"supplies the reference centerline data used to validate the lid-driven cavity result.","marker":"[25]"},{"why":"provides the sparse-data assimilation approach and incompressible Navier-Stokes PINN baseline used for the wake recovery.","marker":"[5]"},{"why":"gives the rigorized convergence of the artificial-compressibility solution to the incompressible weak solution as epsilon tends to zero.","marker":"[17]"},{"why":"documents sensitivity of unsteady cylinder-wake data assimilation to observation location, grounding the paper's sensor-placement finding.","marker":"[22]"}],"fun_headline_variants":["Artificial compressibility: a PINN-friendly flow constraint","AC-PINN: trainable incompressible flow, but wakes need data","Relaxed divergence solves stiff PINN flow training","PINN flow solver: relax the constraint, add sensors for wake","Divergence scales with eps: tune it to train PINNs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that minimizing the mean-squared PDE residual at sampled collocation points is a faithful proxy for solving the equations, so the optimizer finds the physically correct solution rather than a low-residual alternative such as the steady symmetric cylinder wake.","fun_headline_variants_meta":{"raw":{"variants":["Artificial compressibility: a PINN-friendly flow constraint","AC-PINN: trainable incompressible flow, but wakes need data","Relaxed divergence solves stiff PINN flow training","PINN flow solver: relax the constraint, add sensors for wake","Divergence scales with eps: tune it to train PINNs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001084,"raw_usage":{"total_tokens":4566,"prompt_tokens":1011,"completion_tokens":3555,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":627,"completion_tokens_details":{"reasoning_tokens":3467}},"tokens_in":627,"tokens_out":3555,"duration_ms":25192,"temperature":1.0,"reasoning_tokens":3467,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:41:44.336691+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Train the plain forward AC-PINN on the Re=100 cylinder with no data-misfit term and a strong, sustained symmetry-breaking inflow, then measure the wake probe's transverse velocity RMS and spectral peak; the paper predicts collapse to the steady branch with RMS around 3e-3 and no Strouhal peak near 0.17. If the network sheds with Strouhal approximately 0.17 without data, the claim that data-free forward solves cannot reach the instability is falsified.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduces the artificial-compressibility idea from which the relaxed continuity equation is taken."},{"cited_title":"Raissi, P","cited_arxiv_id":null,"evidence_quote":"establishes the physics-informed neural network framework used for all residual minimization."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"documents gradient pathologies from stiff constraints, motivating the local-residual reformulation."},{"cited_title":"Unsupervised simulation of incompressible flows with physics- and equality- constrained artificial neural networks","cited_arxiv_id":"2511.18820","evidence_quote":"reports the same collapse to a steady symmetric branch for a constrained PINN on the cylinder wake."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the reference centerline data used to validate the lid-driven cavity result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the sparse-data assimilation approach and incompressible Navier-Stokes PINN baseline used for the wake recovery."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the rigorized convergence of the artificial-compressibility solution to the incompressible weak solution as epsilon tends to zero."},{"cited_title":"Mons, J.-C","cited_arxiv_id":null,"evidence_quote":"documents sensitivity of unsteady cylinder-wake data assimilation to observation location, grounding the paper's sensor-placement finding."}],"review_version":2}