Trustworthy Koopman Operator Learning: Invariance Diagnostics and Error Bounds

Gustav Conradie, Nicolas Boullé, Jean-Christophe Loiseau, Steven L. Brunton, Matthew J. Colbrook

arXivarxiv/2603.15091Mar 16, 2026math.NAscore 0

Koopman operator theory provides a global linear representation of nonlinear dynamics and underpins many data-driven methods. In practice, however, finite-dimensional feature spaces induced by a user-chosen dictionary are rarely invariant, so closure failures and projection errors lead to spurious eigenvalues, misleading Koopman modes, and overconfident forecasts. This paper addresses a central validation problem in data-driven Koopman methods: how to quantify invariance and projection errors for an arbitrary feature space using only snapshot data, and how to use these diagnostics to produce…

SourceJSON

Comments and reviews

No comments yet.

API

curl /api/v1/records/arxiv/2603.15091
curl -X POST /api/v1/records/arxiv/2603.15091/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2603.15091/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"body": "…", "review": true}'