- Rating
- 1
- Confidence
- 3
- Informativeness
- 1
- Presentation
- 1
- Soundness
- 1
Read review
Paper summary
This paper investigates the "parameter response" of neural operators. The authors define a parameter response Jacobian and a relative response geometry error and use those to organize three decisions: selecting a frozen forward model under a protected parameter shift, selecting which frozen model to invert from sparse sensors, and "repairing" a pretrained predictor with physics, tangent, and rollout-response losses. The supporting "theory" is a pathwise telescoping identity (the fundamental theorem of calculus), a Taylor curvature bound, a Gauss--Newton/Fisher linearization, and the tangent-linear chain rule. Experiments are carried out on Burgers, Darcy and Navier-Stokes equations. The main finding is that a shift-calibrated response score reduces Navier--Stokes OOD field error from to .
Critical strengths
- The evaluation protocol is frozen before the protected outcomes, and the negative results are reported rather than hidden.
- The rank-vs-top-one dissociation on Burgers is a genuinely useful observation: better population ranking does not imply better model selection.
Critical weaknesses
- The headline forward result is confounded, and the missing control is decisive. The response selector is tuned on labeled shifted-development outcomes, while the baseline is a strict in-distribution field-L2 rule. No shift-calibrated field-only control is reported. The winning candidate is also a different family (SC-FNO vs. DeepONet). Thus the abstract's cannot be attributed to response geometry; it conflates using shifted labels, using response features, and switching architectures. This is the paper's only large positive selection result, so the framing does not stand.
- The inverse evidence is null or exploratory. The prespecified sensor-geometry score improves inverse log-parameter RMSE from to but does not survive multiplicity correction. The compact score that looks strong reaches ---numerically identical to the oracle---with after being selected post hoc, and the authors themselves call it exploratory. There is no confirmatory inverse result, and the repair experiment makes inverse recovery worse.
- The statistical and reproducibility record is inadequate for the claims. Tables 2 and 3 report no intervals and no seed replication; the candidates per PDE are individual checkpoints (), not seed-averaged ensembles; and the forward headline rests on a single parent-level sign-flip test over parents (). No architectures, loss weights, optimizer, seed integers, or compute table are given, the relative-OOD denominator is never defined, and the inverse failure fractions appear only in prose. Tables 2--4 are also mislabeled or incomplete (the "Response score" column holds OOD error, Table 3's failure column is ambiguous, Table 4 omits the field-control inverse RMSE), and the custom, unreleased generator makes the numbers externally incomparable.
- Novelty: This is standard sensitivity analysis repackaged as a decision framework, and the experiments do not support the repackaging. Eq. 3 is the fundamental theorem of calculus, Eq. 4 is Taylor's theorem, Eq. 5 is Gauss--Newton/Fisher linearization, and Eq. 10 is the chain rule; the authors concede there is no new theorem. There are also two important uncited works: sTCL (Yang, Lu & Mou, arXiv:2609.32797), which enforces derivative consistency on the fly without offline tangent labels, and PhysGuard (arXiv:2606.16602), which uses a Fisher information matrix to protect physics-critical directions during fine-tuning of a neural operator. Both narrow the repair contribution. What remains is a decision framing, and its central selection claim is confounded.
Questions for authors
-
What happens if you run the same selector with the response features removed, tuned on the same labeled shifted cases? If that field-only selector also approaches , the headline gain is not about response geometry.
-
The response selector picks an SC-FNO while the baseline picks a DeepONet. What is the top-one result within each family? Without that comparison, the gain may be an architecture switch, not a response-geometry effect.
-
Four non-isolating repair arms, including a random-reference arm, pass the same corrected field test, the winner is not equal-compute, and the compact inverse score is exploratory, equals the oracle at , and has . What claim remains that is both confirmatory and specific to response geometry?