Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Scheduler4861
Format: Recurso digital
Sprache:
Veröffentlicht: Zenodo 2026
Online-Zugang:https://doi.org/10.5281/zenodo.20001818
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866901959848493056
author Scheduler4861
author_facet Scheduler4861
contents <p>Abstract</p> <p>We investigate how survival conditioning alters Koopman operator representations in XOR-based causal dynamics using weighted Extended Dynamic Mode Decomposition (EDMD).</p> <p>We show that survival conditioning induces a deformation of the underlying probability measure without modifying the dynamics. This deformation systematically breaks operator normality: both Henrici’s departure and the numerical abscissa increase monotonically with conditioning strength. These non-normality measures are strongly correlated with imaginary spectral broadening (r ≈ 0.85–0.89).</p> <p>Pseudospectral analysis reveals resolvent amplification and spectral bulging under conditioning. A null model that preserves marginal distributions but destroys temporal correlations removes these effects, indicating that the phenomenon arises from causal temporal structure.</p> <p>At the operator level, the mechanism is traced to non-commutation of weighted covariance operators, yielding a non-self-adjoint Koopman operator. The results establish a structural link between measure deformation, operator non-normality, and spectral deformation.</p> <p>This work provides a reproducible empirical foundation for a general operator-theoretic principle: probabilistic conditioning can induce non-normal Koopman dynamics through measure deformation.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_20001818
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle 月詠 Tsukuyomi Core Ver.3.0.0 Non-Normal Koopman Operators Induced by Survival Conditioning
Scheduler4861
<p>Abstract</p> <p>We investigate how survival conditioning alters Koopman operator representations in XOR-based causal dynamics using weighted Extended Dynamic Mode Decomposition (EDMD).</p> <p>We show that survival conditioning induces a deformation of the underlying probability measure without modifying the dynamics. This deformation systematically breaks operator normality: both Henrici’s departure and the numerical abscissa increase monotonically with conditioning strength. These non-normality measures are strongly correlated with imaginary spectral broadening (r ≈ 0.85–0.89).</p> <p>Pseudospectral analysis reveals resolvent amplification and spectral bulging under conditioning. A null model that preserves marginal distributions but destroys temporal correlations removes these effects, indicating that the phenomenon arises from causal temporal structure.</p> <p>At the operator level, the mechanism is traced to non-commutation of weighted covariance operators, yielding a non-self-adjoint Koopman operator. The results establish a structural link between measure deformation, operator non-normality, and spectral deformation.</p> <p>This work provides a reproducible empirical foundation for a general operator-theoretic principle: probabilistic conditioning can induce non-normal Koopman dynamics through measure deformation.</p>
title 月詠 Tsukuyomi Core Ver.3.0.0 Non-Normal Koopman Operators Induced by Survival Conditioning
url https://doi.org/10.5281/zenodo.20001818