A note on large deviations for unbounded observables
Andrew Torok, Matthew Nicol

TL;DR
This paper investigates large deviations for unbounded observables in uniformly expanding systems, showing that uniform expansion alone does not guarantee a rate function and providing specific examples with various decay behaviors.
Contribution
It demonstrates that uniform expansion does not ensure a rate function for unbounded observables and provides detailed examples with different decay rates and behaviors.
Findings
Uniform expansion does not imply the existence of a rate function for unbounded observables.
Examples of unbounded observables with exponential decay of autocorrelations and stretched exponential large deviations.
A classical example of a bounded observable with no rate function is a degenerate coboundary.
Abstract
We consider exponential large deviations estimates for unbounded observables on uniformly expanding dynamical systems. We show that uniform expansion does not imply the existence of a rate function for unbounded observables no matter the tail behavior of the cumulative distribution function. We give examples of unbounded observables with exponential decay of autocorrelations, exponential decay under the transfer operator in each , , and strictly stretched exponential large deviation. For observables of form , periodic, on uniformly expanding systems we give the precise stretched exponential decay rate. We also show that a classical example in the literature of a bounded observable with exponential decay of autocorrelations yet with no rate function is degenerate as the observable is a coboundary.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsStability and Controllability of Differential Equations · Stochastic processes and financial applications · Gene Regulatory Network Analysis
