{"id":"e6c0800e-c705-4f34-8187-2ff0076275c2","arxiv_id":"2606.04125","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Systematic benchmark of PINN architectures on 1D stiff PNP system finds BRDR loss weighting competitive with NTK at lower wall-clock time.","lead":"The paper benchmarks eleven physics-informed neural network configurations across four strategy groups on a one-dimensional stiff Poisson-Nernst-Planck model of a lithium symmetric cell. It identifies the balanced residual decay rate scheme as matching neural tangent kernel accuracy on concentrations while using less compute time, with open-source code released.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.3","headline":"Benchmark limited to 1D lithium symmetric cell; BRDR preference may not hold in 2D/3D or varied stiffness regimes","rationale":"The reader's weakest assumption already isolates the representativeness issue; the 1D restriction is the single most load-bearing limitation on the headline claim that BRDR is preferable under compute constraints.","tokens_in":1711,"tokens_out":306,"duration_ms":13577,"concrete_test":"Re-run the identical eleven-configuration suite on a 2D axisymmetric lithium-cell geometry with the same FVM reference solution and the same PhysicsNeMo Sym implementation; if the concentration RMSE gap between BRDR and NTK exceeds 10 % or the wall-clock ordering reverses, the compute-constraint preference does not generalize.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim ranks BRDR above NTK for concentration-field RMSE at lower wall-clock cost on the basis of eleven configurations tested exclusively on a 1D physically parametrized PNP model. For the claim to support a general recommendation, the relative performance of loss-weighting and multi-scale strategies must be insensitive to dimensionality and to the specific boundary-layer geometry of the symmetric cell. In higher dimensions the electric double layer becomes a surface rather than points, the number of collocation points required for resolution grows, and the interaction between Poisson and Nernst-Planck residuals changes; nothing in the reported experiments tests whether BRDR retains its reported advantage under those conditions.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper conducts the first systematic benchmark of eleven PINN configurations grouped into four strategy groups for solving the stiff one-dimensional Poisson-Nernst-Planck equations modeling a lithium symmetric cell. Implemented in NVIDIA PhysicsNeMo Sym and validated against a finite volume method reference, the study concludes that the balanced residual decay rate (BRDR) scheme achieves performance comparable to the Neural Tangent Kernel (NTK) approach for concentration fields while reducing mean wall-clock time, making it preferable under compute constraints. Loss landscape analysis supports the RMSE rankings, and the code is released openly.","tokens_in":1840,"tokens_out":424,"duration_ms":17444,"significance":"If the empirical findings hold, this work provides valuable guidance on loss-weighting and multi-scale strategies for PINNs applied to stiff coupled PDEs like PNP systems. The open-source PhysicsNeMo Sym implementation is a clear strength, enabling reproducibility and extension to other computational mechanics problems.","major_comments":[{"comment":"Abstract: The abstract states that RMSE spans architectures and that BRDR matches NTK while lowering wall-clock time, but supplies no numerical values, error bars, data exclusion criteria, or validation details against the FVM reference. This absence makes it impossible to assess the magnitude or statistical reliability of the claimed match.","section":"Abstract"},{"comment":"Abstract: The recommendation that BRDR is the preferable strategy under compute constraints rests entirely on results from a one-dimensional physically parametrized lithium symmetric cell. No experiments or discussion address whether the relative performance of BRDR versus NTK persists when the electric double layer becomes a surface in 2D/3D or when stiffness regimes and collocation requirements change.","section":"Abstract"}],"minor_comments":[{"comment":"The four strategy groups and eleven configurations would benefit from an explicit summary table listing each architecture, its loss-weighting or multi-scale component, and key hyperparameters.","section":null}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive comments on our manuscript. We address each major comment below.","responses":[{"response":"The abstract is intended as a concise overview. The manuscript provides full numerical RMSE values, standard deviations across runs, explicit comparison criteria against the FVM reference, and loss-landscape diagnostics in the results section and supplementary tables. To improve standalone readability of the abstract we will insert representative quantitative values (e.g., mean RMSE for concentration fields under BRDR and NTK) together with a brief statement on the validation protocol.","revision_made":"yes","referee_comment":"[Abstract] Abstract: The abstract states that RMSE spans architectures and that BRDR matches NTK while lowering wall-clock time, but supplies no numerical values, error bars, data exclusion criteria, or validation details against the FVM reference. This absence makes it impossible to assess the magnitude or statistical reliability of the claimed match."},{"response":"The study is deliberately scoped to a canonical one-dimensional stiff PNP problem to enable a controlled, systematic comparison of eleven architectures. The manuscript makes no claim of dimensional generality; the recommendation is explicitly tied to the 1-D lithium-symmetric-cell setting under the reported stiffness and collocation conditions. Extending the benchmark to 2-D/3-D geometries constitutes a substantial separate investigation that lies outside the present scope.","revision_made":"no","referee_comment":"[Abstract] Abstract: The recommendation that BRDR is the preferable strategy under compute constraints rests entirely on results from a one-dimensional physically parametrized lithium symmetric cell. No experiments or discussion address whether the relative performance of BRDR versus NTK persists when the electric double layer becomes a surface in 2D/3D or when stiffness regimes and collocation requirements change."}],"tokens_in":1336,"tokens_out":388,"duration_ms":16394,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main takeaway is that this paper runs the first systematic benchmark of eleven PINN configurations on a one-dimensional stiff Poisson-Nernst-Planck system for a lithium symmetric cell. It concludes that the balanced residual decay rate (BRDR) scheme matches the performance of the neural tangent kernel (NTK) approach on concentration fields while cutting down on wall-clock time, making it preferable when resources are limited.\n\nThe paper does a few things right. It organizes the configurations into four strategy groups and tests them all data-free inside the NVIDIA PhysicsNeMo Sym framework. Validation comes from an independent finite volume method reference, which is good for checking the PINN results. They also examine the loss landscape geometry to support the RMSE rankings. Releasing the open-source implementation is a practical move that lets others reuse the setup on similar stiff coupled PDE problems in mechanics.\n\nOn the new side, the specific ranking and the identification of BRDR as a compute-friendly option for this PNP case do not appear in the referenced prior work. The abstract positions it as the first such data-free benchmark on this system.\n\nThe soft spots are mostly around scope. All tests stay in one dimension on a single physically parametrized model. The stress test note is on point here: in two or three dimensions the double layer turns into a surface, the number of points needed grows, and the residual balance might shift. We have no data on whether BRDR keeps its reported advantage outside this 1D symmetric cell. The abstract talks about RMSE spanning architectures and BRDR matching NTK but does not include the actual numbers, error bars, or details on how many runs went into the mean wall-clock times. That makes it tougher to gauge how robust the preference is.\n\nThis paper is aimed at researchers using PINNs for electrodiffusion or battery modeling who need practical advice on loss weighting and multi-scale handling. A reader looking for guidance on stiff PNP problems in 1D would get value from the comparisons and the code. It deserves a serious referee because the benchmark is organized, the validation is external, and the code release supports reproducibility, even if the conclusions stay tied to the 1D case.","headline":"The paper runs the first systematic benchmark of eleven PINN configs on a 1D lithium-cell PNP model and finds BRDR matches NTK accuracy on concentrations at lower wall-clock cost.","tokens_in":2336,"tokens_out":525,"would_cite":false,"duration_ms":19785,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"The balanced residual decay rate scheme matches neural tangent kernel accuracy on concentration fields for stiff Poisson-Nernst-Planck problems while cutting wall-clock time.","keywords":["physics-informed neural networks","Poisson-Nernst-Planck","stiff systems","adaptive loss weighting","neural tangent kernel","benchmark","lithium symmetric cell","electric double layer"],"falsifier":"A repeat of the eleven-configuration benchmark on a two-dimensional PNP geometry or a materially different physical parameter set in which the BRDR scheme no longer matches NTK accuracy on concentrations or loses its wall-clock advantage.","tokens_in":2602,"feed_emoji":"🔋","tokens_out":718,"duration_ms":19205,"temperature":0.7,"pith_summary":"The paper establishes the first systematic benchmark of eleven physics-informed neural network configurations grouped into four strategy families on a one-dimensional Poisson-Nernst-Planck model of a lithium symmetric cell. It reports that the balanced residual decay rate scheme reaches neural tangent kernel performance on concentration profiles yet requires less mean computation time, with loss-landscape geometry aligning with the root-mean-square error ordering. A sympathetic reader would care because the Poisson-Nernst-Planck system is a canonical stiff coupled problem whose extreme coefficient ratios and thin electric-double-layer boundary layers defeat both conventional meshes and standard neural training due to spectral bias and multi-task loss imbalance. The work supplies an open-source implementation inside PhysicsNeMo Sym for reuse on similar stiff electrokinetic systems.","feed_headline":"BRDR loss scheme matches NTK on stiff PNP with less compute time","feed_subtitle":"Benchmark of eleven PINN configurations on one-dimensional lithium-cell model shows balanced residual decay rate as efficient option under t","key_machinery":"The balanced residual decay rate (BRDR) scheme, which dynamically reweights individual loss terms according to the observed decay rates of their residuals to counteract multi-task imbalance during training of physics-informed networks on stiff coupled PDEs.","core_discovery":"Among the tested architectures the balanced residual decay rate scheme matches neural tangent kernel performance for concentration fields while reducing mean wall-clock time, making it the preferable strategy under compute constraints; root-mean-square errors vary across the eleven configurations and loss-landscape geometry corroborates the ranking.","pith_inferences":["The time advantage of BRDR may extend to other multi-physics stiff systems whose loss terms decay at mismatched rates.","A two-dimensional or three-dimensional version of the same benchmark would test whether the observed ranking survives increased spatial complexity.","Open release of the code lowers the threshold for testing PINNs on electrokinetic transport in batteries, membranes, and biological ion channels."],"forward_implications":["BRDR becomes the strategy of choice when wall-clock time is the binding constraint for concentration-field accuracy.","Loss-landscape geometry supplies an independent diagnostic that tracks RMSE rankings across architectures.","The released PhysicsNeMo Sym implementation can be applied directly to other stiff coupled PDE problems in computational mechanics.","Adaptive loss-weighting strategies mitigate the multi-task imbalance that otherwise limits PINN accuracy on stiff PNP systems."],"fun_headline_variants":["BRDR matches NTK on stiff PNP while reducing wall clock time","Benchmark of eleven PINNs on PNP favors BRDR for efficiency","BRDR scheme matches NTK performance on concentration fields","Loss landscape corroborates BRDR ranking in PNP tests"],"cache_read_input_tokens":64,"weakest_assumption_plain":"The eleven PINN configurations organized into four strategy groups, the one-dimensional physically parametrised PNP model for a lithium symmetric cell, and the finite volume method reference are representative enough to rank architectures for general stiff PNP problems.","fun_headline_variants_meta":{"raw":{"variants":["BRDR matches NTK on stiff PNP while reducing wall clock time","Benchmark of eleven PINNs on PNP favors BRDR for efficiency","BRDR scheme matches NTK performance on concentration fields","Loss landscape corroborates BRDR ranking in PNP tests"]},"model":"grok-4.3","cost_usd":0.007517,"raw_usage":{"total_tokens":3426,"prompt_tokens":623,"num_sources_used":0,"completion_tokens":63,"cost_in_usd_ticks":75174500,"prompt_tokens_details":{"text_tokens":623,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2740,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":623,"tokens_out":63,"duration_ms":18400,"temperature":1.0,"reasoning_tokens":2740,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-28T07:06:34.057158+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A repeat of the eleven-configuration benchmark on a two-dimensional PNP geometry or a materially different physical parameter set in which the BRDR scheme no longer matches NTK accuracy on concentrations or loses its wall-clock advantage.","supporting_citations":[],"review_version":1}