Distilled equations predicted instability in simulated multi-planet systems
Source check: Source checked. Not independent scientific confirmation.
Finding: Researchers distilled a neural network trained on three-planet simulations into compact equations for orbital instability time. Tested on previously simulated five-planet systems, the equations captured changes of orders of magnitude near strong orbital resonances. A systematic error remained for closely spaced planets.
Technology: Sparse neural feature learning; PySR symbolic regression
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- The network learned sparse combinations of orbital elements. PySR converted its predicted instability times into closed-form expressions over those learned variables.
- Test and result
- The equation used sparse statistics from short numerical integrations. It was tested on held-out three-planet simulations and on an existing five-planet grid under a neighboring-trio approximation.
- Limitation / still unknown
- The five-planet grid used equal masses and circular coplanar orbits. Results are approximate predictions of simulated dynamics, with compressed prediction ranges and no observational validation.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://arxiv.org/abs/2610.35617v1 · https://arxiv.org/html/2610.35617v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 21 Nov 2026. Review scope: Model equations, Results, Figures 2–4, Appendices A–C and physical interpretation.
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.