SPIRAL-PO: Symbolic Identification of Partially Observed Nonlinear Dynamics with Application to Rotating Machinery
arXiv:2608.00466
2026
Dynamics
2 ideas extracted · analyzed Aug 31, 2026
What the math gives to ML
The paper offers a constructive mechanism for learning partially observed dynamics: separate measured-coordinate physics from a hidden-state coupling residual, fit that residual, project it onto a constrained symbolic library, and admit terms only after sequential statistical tests. Its strongest transferable asset is the identifiability logic: hidden effects can be recovered only when the input or operating-condition trajectory makes the candidate-library feature matrix sufficiently rich. A practical neural-network transfer is a latent neural ODE or state-space model whose observable residual is constrained by a physics library and periodically pruned by statistical gating, with excitation richness used as a data-collection or curriculum criterion.
Ideas from this paper
△ Mechanism confirmed, baseline not beaten
2026
Augment a latent neural state-space model with an observable-coordinate residual that is first learned flexibly and then projected onto a constrained library of interpretable coupling terms. Train or collect data only after checking that the trajectory sufficiently excites the candidate terms; this prevents a latent model from fitting arbitrary hidden-state effects that are unidentifiable from the observations.
Useful8/10
Difficulty6/10
Novelty7/10
✗ Mechanism failed
2026
Use a two-stage residual augmentation loop: first let a residual network explain model mismatch, then project its learned vector field onto a physically constrained candidate library and replace the flexible residual with the accepted sparse terms. This turns an unconstrained neural correction into a low-complexity dynamical law that is easier to roll out over long horizons and can expose unsupported hidden-state explanations.
Useful7/10
Difficulty5/10
Novelty6/10