Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userAn inequality connecting entropy distance, Fisher Information and large deviations
201
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
In this paper we introduce a new generalisation of the relative Fisher Information for Markov jump processes on a finite or countable state space, and prove an inequality which connects this object with the relative entropy and a large deviation rate functional. In addition to possessing various favourable properties, we show that this generalised Fisher Information converges to the classical Fisher Information in an appropriate limit. We then use this generalised Fisher Information and the aforementioned inequality to qualitatively study coarse-graining problems for jump processes on discrete spaces.
rate research
Read More
This paper is a strongly geometrical approach to the Fisher distance, which is a measure of dissimilarity between two probability distribution functions. The Fisher distance, as well as other divergence measures, are also used in many applications to establish a proper data average. The main purpose is to widen the range of possible interpretations and relations of the Fisher distance and its associated geometry for the prospective applications. It focuses on statistical models of the normal probability distribution functions and takes advantage of the connection with the classical hyperbolic geometry to derive closed forms for the Fisher distance in several cases. Connections with the well-known Kullback-Leibler divergence measure are also devised.
We consider the inequality $f geqslant fstar f$ for real integrable functions on $d$ dimensional Euclidean space where $fstar f$ denotes the convolution of $f$ with itself. We show that all such functions $f$ are non-negative, which is not the case for the same inequality in $L^p$ for any $1 < p leqslant 2$, for which the convolution is defined. We also show that all integrable solutions $f$ satisfy $int f(x){rm d}x leqslant tfrac12$. Moreover, if $int f(x){rm d}x = tfrac12$, then $f$ must decay fairly slowly: $int |x| f(x){rm d}x = infty$, and this is sharp since for all $r< 1$, there are solutions with $int f(x){rm d}x = tfrac12$ and $int |x|^r f(x){rm d}x <infty$. However, if $int f(x){rm d}x = : a < tfrac12$, the decay at infinity can be much more rapid: we show that for all $a<tfrac12$, there are solutions such that for some $epsilon>0$, $int e^{epsilon|x|}f(x){rm d}x < infty$.
For hyperbolic flows $varphi_t$ we examine the Gibbs measure of points $w$ for which $$int_0^T G(varphi_t w) dt - a T in (- e^{-epsilon n}, e^{- epsilon n})$$ as $n to infty$ and $T geq n$, provided $epsilon > 0$ is sufficiently small. This is similar to local central limit theorems. The fact that the interval $(- e^{-epsilon n}, e^{- epsilon n})$ is exponentially shrinking as $n to infty$ leads to several difficulties. Under some geometric assumptions we establish a sharp large deviation result with leading term $C(a) epsilon_n e^{gamma(a) T}$ and rate function $gamma(a) leq 0.$ The proof is based on the spectral estimates for the iterations of the Ruelle operators with two complex parameters and on a new Tauberian theorem for sequence of functions $g_n(t)$ having an asymptotic as $ n to infty$ and $t geq n.$
In this note, we study large deviations of the number $mathbf{N}$ of intercalates ($2times2$ combinatorial subsquares which are themselves Latin squares) in a random $ntimes n$ Latin square. In particular, for constant $delta>0$ we prove that $Pr(mathbf{N}le(1-delta)n^{2}/4)leexp(-Omega(n^{2}))$ and $Pr(mathbf{N}ge(1+delta)n^{2}/4)leexp(-Omega(n^{4/3}(log n)^{2/3}))$, both of which are sharp up to logarithmic factors in their exponents. As a consequence, we deduce that a typical order-$n$ Latin square has $(1+o(1))n^{2}/4$ intercalates, matching a lower bound due to Kwan and Sudakov and resolving an old conjecture of McKay and Wanless.
The paper studies the sampling discretization problem for integral norms on subspaces of $L^p(mu)$. Several close to optimal results are obtained on subspaces for which certain Nikolskii-type inequality is valid. The problem of norms discretization is connected with the probabilistic question about the approximation with high probability of marginals of a high dimensional random vector by sampling. As a byproduct of our approach we refine the result of O. Gu$acute{e}$don and M. Rudelson concerning the approximation of marginals. In particular, the obtained improvement recovers a theorem of J. Bourgain, J. Lindenstrauss, and V. Milman concerning embeddings of finite dimensional subspaces of $L^p[0, 1]$ into $ell_p^m$. The proofs in the paper use the recent developments of the chaining technique by R. van Handel.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
