Koopman Invariants as Drivers of Emergent Time-Series Clustering in Joint-Embedding Predictive Architectures

Pablo Ruiz-Morales; Dries Vanoost; Davy Pissoort; Mathias Verbeke

AAAI 2026technical0 citations

Koopman Invariants as Drivers of Emergent Time-Series Clustering in Joint-Embedding Predictive Architectures

Pablo Ruiz-Morales, Dries Vanoost, Davy Pissoort, Mathias Verbeke

Abstract

Joint-Embedding Predictive Architectures (JEPAs), a powerful class of self-supervised models, exhibit an unexplained ability to cluster time-series data by their underlying dynamical regimes. We propose a novel theoretical explanation for this phenomenon, hypothesizing that JEPA

BibTeX

@inproceedings{aaai2026_koopmaninvariant,
  title = {Koopman Invariants as Drivers of Emergent Time-Series Clustering in Joint-Embedding Predictive Architectures},
  author = {Pablo Ruiz-Morales and Dries Vanoost and Davy Pissoort and Mathias Verbeke},
  booktitle = {AAAI 2026},
  year = {2026}
}