Recurrence and repetition times in the case of a stretched exponential growth

06/26/2023
by   Łukasz Dębowski, et al.
0

By an analogy to the duality between the recurrence time and the longest match length, we introduce a quantity dual to the maximal repetition length, which we call the repetition time. Using the generalized Kac lemma for successive recurrence times by Chen Moy, we sandwich the repetition time in terms of min-entropies with no or relatively short conditioning. The sole assumption is stationarity and ergodicity. The proof is surprisingly short and the claim is fully general in contrast to earlier approaches by Szpankowski and by Dębowski. We discuss the analogy of this result with the Wyner-Ziv/Ornstein-Weiss theorem, which sandwiches the recurrence time in terms of Shannon entropies. We formulate the respective sandwich bounds in a way that applies also to the case of stretched exponential growth observed empirically for natural language.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
05/14/2019

Quantum Complexity of Time Evolution with Chaotic Hamiltonians

We study the quantum complexity of time evolution in large-N chaotic sys...
research
07/26/2021

Tsallis and Rényi deformations linked via a new λ-duality

Tsallis and Rényi entropies, which are monotone transformations of such ...
research
09/11/2020

Linear Shannon Capacity of Cayley Graphs

The Shannon capacity of a graph is a fundamental quantity in zero-error ...
research
09/07/2023

LRCs: Duality, LP Bounds, and Field Size

We develop a duality theory of locally recoverable codes (LRCs) and appl...
research
05/16/2022

Natural Language Specifications in Proof Assistants

Interactive proof assistants are computer programs carefully constructed...
research
03/01/2023

E-values for k-Sample Tests With Exponential Families

We develop and compare e-variables for testing whether k samples of data...
research
12/22/2017

A short variational proof of equivalence between policy gradients and soft Q learning

Two main families of reinforcement learning algorithms, Q-learning and p...

Please sign up or login with your details

Forgot password? Click here to reset