Large-deviation theory and coverage in mobile phone networks

Size: px
Start display at page:

Download "Large-deviation theory and coverage in mobile phone networks"

Transcription

1 Weierstrass Institute for Applied Analysis and Stochastics Large-deviation theory and coverage in mobile phone networks Paul Keeler, Weierstrass Institute for Applied Analysis and Stochastics, Berlin joint work with Christian Hirsch, Benedikt Jahnel, and Robert Patterson. Second half of talk written by Christian Hirsch. Melbourne, August 17th, 2015

2 1 Introduction to large deviations theory 2 SINR definition and model description 3 LDP for averages of connectable receivers in large domains 4 LDP for connectable receivers for small SINR threshold 5 Importance sampling for rare-event probability estimation LDT in phone networks August 17th, 2015 Page 2

3 Large deviations theory (LDT): overview Simple idea: for sample means, the probability distributions of large deviations away from the (true) mean often contain exponential terms LDT bridges the law of large numbers and central limit theorem, which is concerned with small deviations around the mean Statistical physics can often be reformulated as applications of large deviation theory acccording to Touchette (2009) so some ideas go back to Boltzmann. Cramér (1938) derived an early result of LDT known as Cramér s theorem. Connection between LDT and entropy by Sanov. Much theory developed (independently) by Donsker and Varadhan (in the US) and Freidlin and Wentzell (in the USSR) say Dembo and Zeitouni (2009) Mathematics of LDT soon becomes quite abstract and technical S. R. Srinivasa Varadhan, a key figure, won the 2007 Abel prize. Analysis and topology knowledge required for a rigorous treatment, leading to a jungle of deltas and epsilons to establish results Key ideas can be illustrated with well-behaved random variables LDT in phone networks August 17th, 2015 Page 2

4 Simple examples with random variables Collection of i.i.d. random variables {T i } and their sample or empirical mean S n = 1 n T i n If T n are Gaussian with mean µ and standard deviation σ, then the sample mean s probability density p(s) behaves asymptotically (of course, the S n is i=1 another Gaussian, but we neglect the subdominant n term here) p(s) e nj(s) (s µ)2, J(s) = 2σ 2, s R where we make the symbol more clear soon. If T n are exponential with mean 1/µ then the sample mean s probability density p(s) behaves asymptotically (S n is a gamma/erlang random variable) p(s) e nj(s), J(s) = s µ 1 log s µ, s > 0 Rate function of binomial variable with parameter p is the relative entropy of two coins with success probability s and p ie J(s) = s log( s (1 s) p ) + (1 s) log (1 p). J is called the rate function both positive and convex in these examples LDT in phone networks August 17th, 2015 Page 3

5 Large deviation principle (lite or diet version) To show that such exponential behaviour exists is to show a large deviation principle (LDP) exists A n is a random variable indexed by parameter n, assumed to be large P(A n B) is the probability that A n takes on a value in some set (Borel) B Definition (Large deviation principle greatly simplified) P(A n B) satisfies a large deviation principle with rate J B if the following limit exists 1 lim n n log P(A n B) = J B. This is what we mean by P(A n B) e nj B, or with small-o notation log P(A n B) = nj B + o(n) If the above limit doesn t exist, one can sometimes still derive bounds such as e nj B LDT in phone networks August 17th, 2015 Page 4 P(An B) e nj + B

6 Large deviation principle (more general) Definition (Rate function) For a metric space M, a function J : M [0, ] is called a rate function if it is lower semincontinuous, which means the levels sets {x M : J(x) a) are closed for any a 0. Moreover, a rate function is called a good rate function if the levels sets are compact for any a 0. Definition (Large deviation principle) A sequence of random variables T 1, T 2,... (with values in a metric space) satisfy a large deviation principle with speed a n and rate function J if, for all Borel sets B M, the following holds lim sup n lim inf n 1 log P(T n B) inf J(x) a n x clb (1) 1 log P(T n B) inf J(x) a n x intb (2) LDT in phone networks August 17th, 2015 Page 5

7 Cramér s theorem For a random variable T, define the cumulant generating function Theorem (Cramér s theorem) φ(λ) := log Ee λt <, λ R (3) Let T 1, T 2,... be i.i.d. random variables with mean µ satisfying (3) and S n be their sample mean. Then for any x > µ, we have 1 lim n n log P(S n > x) = φ (x), where φ (x) is given by the Legendre(-Fenchel) transform of φ, namely φ (x) := sup[λx φ(λ)] λ R An important generalization is the Gärtner-Ellis theorem: if φ(λ) exists and is differentiable for all λ R, then a LDP exists with rate function J(x) = φ (x) LDT in phone networks August 17th, 2015 Page 6

8 Empirical measure and Sanov s theorem Large deviation principles can be established for more descriptive objects than sample means Assume T 1, T 2,... are i.i.d random variables taking values in a finite set T, and define the empirical measure of a sample vector T = (T 1,..., T n ) by L T n (t) := 1 n 1[T i = t] = 1 n δ Ti (t), t T n n i=1 (interpreted as a random element of the space of probability measures on T ) Theorem (Sanov s theorem) Assume T 1, T 2,... as above and denote their distribution by µ. Then the empirical measures L T n satisfy a large deviation principle with speed n and a good rate function J given by the relative entropy (or Kullback-Leibler divergence) J(ν) = h(ν µ) := t T i=1 ν(t) log ν(t) µ(t) LDT in phone networks August 17th, 2015 Page 7

9 Empirical processes and LDP levels Large deviation principles can be applied to a generalization of an empirical measure: the empirical pair measure L T n,2(t, v) := 1 n 1[T i 1 = t, T i = v], (t, v) T T n i=1 ( For symmetry, on sets T 0 := T n ) Infinite-dimensional version of the empirical pair measure is called an empirical process, introduced by Donsker and Varadhan Levels of large deviation principles: Level 1 for sample/empirical means, Level 2 for empirical measures, and Level 3 for empirical processes Large deviation frameworks have been adapted for various random objects eg for marked point processes with Level-3 results by Georgii and Zessin (1993) LDT in phone networks August 17th, 2015 Page 8

10 Applications and importance sampling Applications: Statistical physics (eg Ising model, Curie-Weiss-Potts model) as the number of particles n approaches infinity (ie thermodynamic limit), with a focus on Gibbsian systems Used to study Markov processes and stochastic differential equations Risk and finance (remember Cramér worked as an actuary for many years) Information theory (which can considered as a generalization of certain parts of statistical physics) Speed up stochastic or Monte Carlo simulations by importance sampling Importance sampling: Large deviation principles gives a more formed way to perform stochastic or Monte Carlo simulations Studying a rare event by simulation takes (by definition) a long time. Change the probability measure so the events occurs more often (compare to change of variables for an integral) LDT in phone networks August 17th, 2015 Page 9

11 Signal-to-noise-plus-interference-ratio (SINR) USEFUL SIGNAL RECEIVED POWER SINR = ALL OTHER (ie INTERFERING) SIGNALS RECEIVED POWER + NOISE Network transmitters located at {X i } in R 2 and a user at the origin o. R i is the received power of a signal from a transmitter at X i R 2. SINR in relation to a transmitter at X i SINR(X i, o) = R i. (4) R j j i SINR is motivated by information theory; eg maximize SINR(X i, o) Propagation model: R i = F i,j l( X i X j ), l(r) is a decreasing function eg l(r) = min[1, r α ] where α > 2, Assume F i,j are i.i.d. random variables representing propagation effects such as multi-path fading and/or shadowing ie signals colliding with obstacles. LDT in phone networks August 17th, 2015 Page 10

12 SINR-based network model Homogeneous Poisson point process of transmitters Homogeneous Poisson point process of receivers Connectable receivers associated with each transmitter Goal: exponential decay of probability of unlikely configurations eg large proportion of isolated transmitters LDT in phone networks August 17th, 2015 Page 11

13 SINR-based network model Homogeneous Poisson point process of transmitters Homogeneous Poisson point process of receivers Connectable receivers associated with each transmitter Goal: exponential decay of probability of unlikely configurations eg large proportion of isolated transmitters LDT in phone networks August 17th, 2015 Page 11

14 SINR-based network model Homogeneous Poisson point process of transmitters Homogeneous Poisson point process of receivers Connectable receivers associated with each transmitter Goal: exponential decay of probability of unlikely configurations eg large proportion of isolated transmitters LDT in phone networks August 17th, 2015 Page 11

15 Network model Network participants X = homogeneous Poisson point process of transmitters in R d ; intensity λ T Y = homogeneous Poisson point process of receivers in R d ; intensity λ R SINR model: for X i X and Y j Y, define the SINR as SINR(X i, Y j ) = SINR(X i, Y j, X) = F Xi,Y j l( X i Y j ) w + k i F X k,y j l( X k Y j ) w = constant thermal noise l(r) = min{1, r α } path-loss function with path-loss exponent α > d Compact support {F x,y } x X, y Y = i.i.d. fading variables Compact support Distribution q of Fx,y 1 is globally Lipschitz LDT in phone networks August 17th, 2015 Page 12

16 1 Introduction to large deviations theory 2 SINR definition and model description 3 LDP for averages of connectable receivers in large domains 4 LDP for connectable receivers for small SINR threshold 5 Importance sampling for rare-event probability estimation LDT in phone networks August 17th, 2015 Page 13

17 Reachable receivers For τ > 0, define connectable receivers Y (i) associated with transmitter X i Y (i) = {Y j Y : SINR(X i, Y j ) τ} Possible applications Device-to-device (D2D) networks ie mobile phones communicate directly to each other First step in a multi-hop network Goal: exponential decay of probability that an atypically large proportion of transmitters suffers from some frustration event Y (i) = Isolation #Y (i) 1 At most 1 connectable receiver Y (i) B ε (X i ) Impossibility of long hops LDT in phone networks August 17th, 2015 Page 14

18 Empirical measure of connectable receivers Empirical measure of connectable receivers L n = 1 Λ n X i Λ n δ Y (i) X i, Λ n = [ n/2, n/2] d For example, L n ({ }) = 1 Λ n #{X i Λ n : X i isolated} For Q P θ any jointly stationary point process of transmitters and receivers, let Q ( ) = E#{X i Λ 1 : Y (i) X i } be the Palm mark measure of {(X i, Y (i) X i )} i 1 For a marked point process on Λ n = [ n/2, n/2] d, the specific relative entropy / Kullback-Leibler divergence is 1 ( h(q P) = lim n Λ n E dq Λn P log dq ) Λ n. dp Λn dp Λn LDT in phone networks August 17th, 2015 Page 15

19 Large deviation principle for the empirical measure Theorem (Level-2 result) The random measures {L n } n 1 satisfy an LDP in the τ -topology with speed Λ n and good rate function Loosely speaking, I(Q) = inf h(q P). Q P θ Q =Q P(L n F ) exp( Λ n inf Q F I(Q)). τ -topology (finer than the usual weak topologies) is generated by maps B µ(b), where B is a bounded Borel set. LDPs/importance sampling for interference in (Ganesh and Torrisi, 2008) or (Torrisi and Leonardi, 2013) Proof based on Level-3 LDP for Poisson point processes (Georgii and Zessin, 1993) LDT in phone networks August 17th, 2015 Page 16

20 1 Introduction to large deviations theory 2 SINR definition and model description 3 LDP for averages of connectable receivers in large domains 4 LDP for connectable receivers for small SINR threshold 5 Importance sampling for rare-event probability estimation LDT in phone networks August 17th, 2015 Page 17

21 SINR-based network model Y τ = connectable receivers for a (typical) transmitter located at the origin o Y τ = {Y j Y : SINR(o, Y j ) τ} Asymptotic behavior as technically-dependent parameter τ 0 LDT in phone networks August 17th, 2015 Page 18

22 SINR-based network model Y τ = connectable receivers for a (typical) transmitter located at the origin o Y τ = {Y j Y : SINR(o, Y j ) τ} Asymptotic behavior as technically-dependent parameter τ 0 LDT in phone networks August 17th, 2015 Page 18

23 Large deviation principle for the empirical measure Theorem (Level-1 result) As τ 0 the random measure { τ d α Y τ (τ 1 α B)} B B(Λ1) satisfies an LDP in the weak topology with rate τ d α and good rate function given by { Λ I(µ) = 1 I y (f(y))dy if f = dµ/dx exists, otherwise, where I y (s) = inf Q P θ ( h(q P) + h ( Pois(s) Pois(λR E Q [P(SINR(o, y) 1 X)]) )). Idea: Y τ is Cox point process with random intensity measure B λ R P(SINR(o, y) τ X)dy + (Georgii & Zessin, 1993) + (Dawson & B Gärtner) LDT in phone networks August 17th, 2015 Page 19

24 Exponential decay of isolation probability Corollary ( lim τ d α log P(Y τ = ) = lim τ d α log Eexp τ 0 τ 0 = Λ 1 λ R Λ τ 1/α ) P(SINR(o, y) τ X)dy ( inf h(q P) + λr E Q [P(SINR(o, y) 1 X)] ) dy. Q P θ Variational characterization of rate function intractable analytically and numerically Minimization problem can be solved for class of Poisson point processes importance-sampling algorithms LDT in phone networks August 17th, 2015 Page 20

25 1 Introduction to large deviations theory 2 SINR definition and model description 3 LDP for averages of connectable receivers in large domains 4 LDP for connectable receivers for small SINR threshold 5 Importance sampling for rare-event probability estimation LDT in phone networks August 17th, 2015 Page 21

26 Importance sampling for rare-event probability estimation Problem: estimate isolation probability p τ for small τ by Monte Carlo method Accurate estimation of p τ by naïve MC requires large number of simulation runs Idea: perform simulation under different probability measure Q that makes rare event more likely ( dp ) P(o is isolated) = Q dq (ω)1 o is isolated(ω). Heuristic: optimal Q conditional distribution under rare event Connection to LDPs: use minimizers in rate function α = 4, w = 1, constant fading, constant transmission powers LDT in phone networks August 17th, 2015 Page 22

27 Importance sampling for rare-event probability estimation Problem: estimate isolation probability p τ for small τ by Monte Carlo method Accurate estimation of p τ by naïve MC requires large number of simulation runs Idea: perform simulation under different probability measure Q that makes rare event more likely ( dp ) P(o is isolated) = Q dq (ω)1 o is isolated(ω). Heuristic: optimal Q conditional distribution under rare event Connection to LDPs: use minimizers in rate function α = 4, w = 1, constant fading, constant transmission powers lim τ 1/2 ( log p τ = inf h(q P) + λr Q(SINR(o, y) 1) ) dy. τ 0 Q P θ Λ 1 Optimization only over inhomogeneous Poisson point processes LDT in phone networks August 17th, 2015 Page 22

28 Importance sampling for isolation probability continuous family of 1D-optimization problems for optimal intensity λ opt (r) N = 1, 000, 000 simulation runs Variance reduction by 78% Isolation probability Variance λ( ) λ( ) λ opt ( ) LDT in phone networks August 17th, 2015 Page 23

29 Literature [1] F. Baccelli and B. Błaszczyszyn. Stochastic Geometry and Wireless Networks: Volume 1: Theory. Now Publishers Inc, [2] A. Dembo and O. Zeitouni. Large deviations techniques and applications, volume 38. Springer Science & Business Media, [3] A. J. Ganesh and G. L. Torrisi. Large deviations of the interference in a wireless communication model. IEEE Trans. Inform. Theory, 54(8): , [4] H.-O. Georgii and H. Zessin. Large deviations and the maximum entropy principle for marked point random fields. Probab. Theory Related Fields, 96(2): , [5] C. Hirsch, B. Jahnel, P. Keeler, and R. I. Patterson. Large-deviation principles for connectable receivers in wireless networks. arxiv preprint arxiv: , [6] G. L. Torrisi and E. Leonardi. Simulating the tail of the interference in a Poisson network model. IEEE Trans. Inform. Theory, 59(3): , [7] H. Touchette. The large deviation approach to statistical mechanics. Physics Reports, 478(1):1 69, LDT in phone networks August 17th, 2015 Page 24

30 LDT in phone networks August 17th, 2015 Page 25 [5, 7, 2,?, 1, 4, 6, 3]

Large Deviations Techniques and Applications

Large Deviations Techniques and Applications Amir Dembo Ofer Zeitouni Large Deviations Techniques and Applications Second Edition With 29 Figures Springer Contents Preface to the Second Edition Preface to the First Edition vii ix 1 Introduction 1

More information

1. Principle of large deviations

1. Principle of large deviations 1. Principle of large deviations This section will provide a short introduction to the Law of large deviations. The basic principle behind this theory centers on the observation that certain functions

More information

Brownian intersection local times

Brownian intersection local times Weierstrass Institute for Applied Analysis and Stochastics Brownian intersection local times Wolfgang König (TU Berlin and WIAS Berlin) [K.&M., Ann. Probab. 2002], [K.&M., Trans. Amer. Math. Soc. 2006]

More information

Stronger wireless signals appear more Poisson

Stronger wireless signals appear more Poisson 1 Stronger wireless signals appear more Poisson Paul Keeler, Nathan Ross, Aihua Xia and Bartłomiej Błaszczyszyn Abstract arxiv:1604.02986v3 [cs.ni] 20 Jun 2016 Keeler, Ross and Xia [1] recently derived

More information

Théorie des grandes déviations: Des mathématiques à la physique

Théorie des grandes déviations: Des mathématiques à la physique Théorie des grandes déviations: Des mathématiques à la physique Hugo Touchette National Institute for Theoretical Physics (NITheP) Stellenbosch, Afrique du Sud CERMICS, École des Ponts Paris, France 30

More information

Gärtner-Ellis Theorem and applications.

Gärtner-Ellis Theorem and applications. Gärtner-Ellis Theorem and applications. Elena Kosygina July 25, 208 In this lecture we turn to the non-i.i.d. case and discuss Gärtner-Ellis theorem. As an application, we study Curie-Weiss model with

More information

Notes on Large Deviations in Economics and Finance. Noah Williams

Notes on Large Deviations in Economics and Finance. Noah Williams Notes on Large Deviations in Economics and Finance Noah Williams Princeton University and NBER http://www.princeton.edu/ noahw Notes on Large Deviations 1 Introduction What is large deviation theory? Loosely:

More information

Ferromagnets and the classical Heisenberg model. Kay Kirkpatrick, UIUC

Ferromagnets and the classical Heisenberg model. Kay Kirkpatrick, UIUC Ferromagnets and the classical Heisenberg model Kay Kirkpatrick, UIUC Ferromagnets and the classical Heisenberg model: asymptotics for a mean-field phase transition Kay Kirkpatrick, Urbana-Champaign June

More information

Entropy and Large Deviations

Entropy and Large Deviations Entropy and Large Deviations p. 1/32 Entropy and Large Deviations S.R.S. Varadhan Courant Institute, NYU Michigan State Universiy East Lansing March 31, 2015 Entropy comes up in many different contexts.

More information

A Note On Large Deviation Theory and Beyond

A Note On Large Deviation Theory and Beyond A Note On Large Deviation Theory and Beyond Jin Feng In this set of notes, we will develop and explain a whole mathematical theory which can be highly summarized through one simple observation ( ) lim

More information

Large deviation theory and applications

Large deviation theory and applications Large deviation theory and applications Peter Mörters November 0, 2008 Abstract Large deviation theory deals with the decay of the probability of increasingly unlikely events. It is one of the key techniques

More information

Large deviations for random projections of l p balls

Large deviations for random projections of l p balls 1/32 Large deviations for random projections of l p balls Nina Gantert CRM, september 5, 2016 Goal: Understanding random projections of high-dimensional convex sets. 2/32 2/32 Outline Goal: Understanding

More information

arxiv: v3 [cond-mat.stat-mech] 29 Feb 2012

arxiv: v3 [cond-mat.stat-mech] 29 Feb 2012 A basic introduction to large deviations: Theory, applications, simulations arxiv:1106.4146v3 [cond-mat.stat-mech] 29 Feb 2012 Hugo Touchette School of Mathematical Sciences, Queen Mary, University of

More information

A large deviation principle for a RWRC in a box

A large deviation principle for a RWRC in a box A large deviation principle for a RWRC in a box 7th Cornell Probability Summer School Michele Salvi TU Berlin July 12, 2011 Michele Salvi (TU Berlin) An LDP for a RWRC in a nite box July 12, 2011 1 / 15

More information

Information Theory and Hypothesis Testing

Information Theory and Hypothesis Testing Summer School on Game Theory and Telecommunications Campione, 7-12 September, 2014 Information Theory and Hypothesis Testing Mauro Barni University of Siena September 8 Review of some basic results linking

More information

Iterative Markov Chain Monte Carlo Computation of Reference Priors and Minimax Risk

Iterative Markov Chain Monte Carlo Computation of Reference Priors and Minimax Risk Iterative Markov Chain Monte Carlo Computation of Reference Priors and Minimax Risk John Lafferty School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 lafferty@cs.cmu.edu Abstract

More information

Global Optimization, Generalized Canonical Ensembles, and Universal Ensemble Equivalence

Global Optimization, Generalized Canonical Ensembles, and Universal Ensemble Equivalence Richard S. Ellis, Generalized Canonical Ensembles 1 Global Optimization, Generalized Canonical Ensembles, and Universal Ensemble Equivalence Richard S. Ellis Department of Mathematics and Statistics University

More information

Generalized Neyman Pearson optimality of empirical likelihood for testing parameter hypotheses

Generalized Neyman Pearson optimality of empirical likelihood for testing parameter hypotheses Ann Inst Stat Math (2009) 61:773 787 DOI 10.1007/s10463-008-0172-6 Generalized Neyman Pearson optimality of empirical likelihood for testing parameter hypotheses Taisuke Otsu Received: 1 June 2007 / Revised:

More information

Large Deviations for Weakly Dependent Sequences: The Gärtner-Ellis Theorem

Large Deviations for Weakly Dependent Sequences: The Gärtner-Ellis Theorem Chapter 34 Large Deviations for Weakly Dependent Sequences: The Gärtner-Ellis Theorem This chapter proves the Gärtner-Ellis theorem, establishing an LDP for not-too-dependent processes taking values in

More information

The large deviation principle for the Erdős-Rényi random graph

The large deviation principle for the Erdős-Rényi random graph The large deviation principle for the Erdős-Rényi random graph (Courant Institute, NYU) joint work with S. R. S. Varadhan Main objective: how to count graphs with a given property Only consider finite

More information

General Theory of Large Deviations

General Theory of Large Deviations Chapter 30 General Theory of Large Deviations A family of random variables follows the large deviations principle if the probability of the variables falling into bad sets, representing large deviations

More information

Theorem 2.1 (Caratheodory). A (countably additive) probability measure on a field has an extension. n=1

Theorem 2.1 (Caratheodory). A (countably additive) probability measure on a field has an extension. n=1 Chapter 2 Probability measures 1. Existence Theorem 2.1 (Caratheodory). A (countably additive) probability measure on a field has an extension to the generated σ-field Proof of Theorem 2.1. Let F 0 be

More information

Introduction to Rare Event Simulation

Introduction to Rare Event Simulation Introduction to Rare Event Simulation Brown University: Summer School on Rare Event Simulation Jose Blanchet Columbia University. Department of Statistics, Department of IEOR. Blanchet (Columbia) 1 / 31

More information

Sample path large deviations of a Gaussian process with stationary increments and regularily varying variance

Sample path large deviations of a Gaussian process with stationary increments and regularily varying variance Sample path large deviations of a Gaussian process with stationary increments and regularily varying variance Tommi Sottinen Department of Mathematics P. O. Box 4 FIN-0004 University of Helsinki Finland

More information

Consistency of the maximum likelihood estimator for general hidden Markov models

Consistency of the maximum likelihood estimator for general hidden Markov models Consistency of the maximum likelihood estimator for general hidden Markov models Jimmy Olsson Centre for Mathematical Sciences Lund University Nordstat 2012 Umeå, Sweden Collaborators Hidden Markov models

More information

Probability for Statistics and Machine Learning

Probability for Statistics and Machine Learning ~Springer Anirban DasGupta Probability for Statistics and Machine Learning Fundamentals and Advanced Topics Contents Suggested Courses with Diffe~ent Themes........................... xix 1 Review of Univariate

More information

Entropy, Inference, and Channel Coding

Entropy, Inference, and Channel Coding Entropy, Inference, and Channel Coding Sean Meyn Department of Electrical and Computer Engineering University of Illinois and the Coordinated Science Laboratory NSF support: ECS 02-17836, ITR 00-85929

More information

Approximations of displacement interpolations by entropic interpolations

Approximations of displacement interpolations by entropic interpolations Approximations of displacement interpolations by entropic interpolations Christian Léonard Université Paris Ouest Mokaplan 10 décembre 2015 Interpolations in P(X ) X : Riemannian manifold (state space)

More information

Delay-Based Connectivity of Wireless Networks

Delay-Based Connectivity of Wireless Networks Delay-Based Connectivity of Wireless Networks Martin Haenggi Abstract Interference in wireless networks causes intricate dependencies between the formation of links. In current graph models of wireless

More information

Path Coupling and Aggregate Path Coupling

Path Coupling and Aggregate Path Coupling Path Coupling and Aggregate Path Coupling 0 Path Coupling and Aggregate Path Coupling Yevgeniy Kovchegov Oregon State University (joint work with Peter T. Otto from Willamette University) Path Coupling

More information

On the Number of Switches in Unbiased Coin-tossing

On the Number of Switches in Unbiased Coin-tossing On the Number of Switches in Unbiased Coin-tossing Wenbo V. Li Department of Mathematical Sciences University of Delaware wli@math.udel.edu January 25, 203 Abstract A biased coin is tossed n times independently

More information

Contraction Principles some applications

Contraction Principles some applications Contraction Principles some applications Fabrice Gamboa (Institut de Mathématiques de Toulouse) 7th of June 2017 Christian and Patrick 59th Birthday Overview Appetizer : Christian and Patrick secret lives

More information

Poisson Processes. Stochastic Processes. Feb UC3M

Poisson Processes. Stochastic Processes. Feb UC3M Poisson Processes Stochastic Processes UC3M Feb. 2012 Exponential random variables A random variable T has exponential distribution with rate λ > 0 if its probability density function can been written

More information

A PARAMETRIC MODEL FOR DISCRETE-VALUED TIME SERIES. 1. Introduction

A PARAMETRIC MODEL FOR DISCRETE-VALUED TIME SERIES. 1. Introduction tm Tatra Mt. Math. Publ. 00 (XXXX), 1 10 A PARAMETRIC MODEL FOR DISCRETE-VALUED TIME SERIES Martin Janžura and Lucie Fialová ABSTRACT. A parametric model for statistical analysis of Markov chains type

More information

Weak quenched limiting distributions of a one-dimensional random walk in a random environment

Weak quenched limiting distributions of a one-dimensional random walk in a random environment Weak quenched limiting distributions of a one-dimensional random walk in a random environment Jonathon Peterson Cornell University Department of Mathematics Joint work with Gennady Samorodnitsky September

More information

Visibility estimates in Euclidean and hyperbolic germ-grain models and line tessellations

Visibility estimates in Euclidean and hyperbolic germ-grain models and line tessellations Visibility estimates in Euclidean and hyperbolic germ-grain models and line tessellations Pierre Calka Lille, Stochastic Geometry Workshop, 30 March 2011 Outline Visibility in the vacancy of the Boolean

More information

ELEMENTS OF PROBABILITY THEORY

ELEMENTS OF PROBABILITY THEORY ELEMENTS OF PROBABILITY THEORY Elements of Probability Theory A collection of subsets of a set Ω is called a σ algebra if it contains Ω and is closed under the operations of taking complements and countable

More information

Lecture 5: Importance sampling and Hamilton-Jacobi equations

Lecture 5: Importance sampling and Hamilton-Jacobi equations Lecture 5: Importance sampling and Hamilton-Jacobi equations Henrik Hult Department of Mathematics KTH Royal Institute of Technology Sweden Summer School on Monte Carlo Methods and Rare Events Brown University,

More information

Rare event simulation for the ruin problem with investments via importance sampling and duality

Rare event simulation for the ruin problem with investments via importance sampling and duality Rare event simulation for the ruin problem with investments via importance sampling and duality Jerey Collamore University of Copenhagen Joint work with Anand Vidyashankar (GMU) and Guoqing Diao (GMU).

More information

Ferromagnets and superconductors. Kay Kirkpatrick, UIUC

Ferromagnets and superconductors. Kay Kirkpatrick, UIUC Ferromagnets and superconductors Kay Kirkpatrick, UIUC Ferromagnet and superconductor models: Phase transitions and asymptotics Kay Kirkpatrick, Urbana-Champaign October 2012 Ferromagnet and superconductor

More information

The Moment Method; Convex Duality; and Large/Medium/Small Deviations

The Moment Method; Convex Duality; and Large/Medium/Small Deviations Stat 928: Statistical Learning Theory Lecture: 5 The Moment Method; Convex Duality; and Large/Medium/Small Deviations Instructor: Sham Kakade The Exponential Inequality and Convex Duality The exponential

More information

Surrogate loss functions, divergences and decentralized detection

Surrogate loss functions, divergences and decentralized detection Surrogate loss functions, divergences and decentralized detection XuanLong Nguyen Department of Electrical Engineering and Computer Science U.C. Berkeley Advisors: Michael Jordan & Martin Wainwright 1

More information

An inverse of Sanov s theorem

An inverse of Sanov s theorem An inverse of Sanov s theorem Ayalvadi Ganesh and Neil O Connell BRIMS, Hewlett-Packard Labs, Bristol Abstract Let X k be a sequence of iid random variables taking values in a finite set, and consider

More information

Ruin, Operational Risk and How Fast Stochastic Processes Mix

Ruin, Operational Risk and How Fast Stochastic Processes Mix Ruin, Operational Risk and How Fast Stochastic Processes Mix Paul Embrechts ETH Zürich Based on joint work with: - Roger Kaufmann (ETH Zürich) - Gennady Samorodnitsky (Cornell University) Basel Committee

More information

LARGE DEVIATIONS FOR DOUBLY INDEXED STOCHASTIC PROCESSES WITH APPLICATIONS TO STATISTICAL MECHANICS

LARGE DEVIATIONS FOR DOUBLY INDEXED STOCHASTIC PROCESSES WITH APPLICATIONS TO STATISTICAL MECHANICS LARGE DEVIATIONS FOR DOUBLY INDEXED STOCHASTIC PROCESSES WITH APPLICATIONS TO STATISTICAL MECHANICS A Dissertation Presented by CHRISTOPHER L. BOUCHER Submitted to the Graduate School of the University

More information

2012 NCTS Workshop on Dynamical Systems

2012 NCTS Workshop on Dynamical Systems Barbara Gentz gentz@math.uni-bielefeld.de http://www.math.uni-bielefeld.de/ gentz 2012 NCTS Workshop on Dynamical Systems National Center for Theoretical Sciences, National Tsing-Hua University Hsinchu,

More information

Lecture 1: Entropy, convexity, and matrix scaling CSE 599S: Entropy optimality, Winter 2016 Instructor: James R. Lee Last updated: January 24, 2016

Lecture 1: Entropy, convexity, and matrix scaling CSE 599S: Entropy optimality, Winter 2016 Instructor: James R. Lee Last updated: January 24, 2016 Lecture 1: Entropy, convexity, and matrix scaling CSE 599S: Entropy optimality, Winter 2016 Instructor: James R. Lee Last updated: January 24, 2016 1 Entropy Since this course is about entropy maximization,

More information

A relative entropy characterization of the growth rate of reward in risk-sensitive control

A relative entropy characterization of the growth rate of reward in risk-sensitive control 1 / 47 A relative entropy characterization of the growth rate of reward in risk-sensitive control Venkat Anantharam EECS Department, University of California, Berkeley (joint work with Vivek Borkar, IIT

More information

TRANSMISSION STRATEGIES FOR SINGLE-DESTINATION WIRELESS NETWORKS

TRANSMISSION STRATEGIES FOR SINGLE-DESTINATION WIRELESS NETWORKS The 20 Military Communications Conference - Track - Waveforms and Signal Processing TRANSMISSION STRATEGIES FOR SINGLE-DESTINATION WIRELESS NETWORKS Gam D. Nguyen, Jeffrey E. Wieselthier 2, Sastry Kompella,

More information

Stability of optimization problems with stochastic dominance constraints

Stability of optimization problems with stochastic dominance constraints Stability of optimization problems with stochastic dominance constraints D. Dentcheva and W. Römisch Stevens Institute of Technology, Hoboken Humboldt-University Berlin www.math.hu-berlin.de/~romisch SIAM

More information

How long before I regain my signal?

How long before I regain my signal? How long before I regain my signal? Tingting Lu, Pei Liu and Shivendra S. Panwar Polytechnic School of Engineering New York University Brooklyn, New York Email: tl984@nyu.edu, peiliu@gmail.com, panwar@catt.poly.edu

More information

EVALUATING SYMMETRIC INFORMATION GAP BETWEEN DYNAMICAL SYSTEMS USING PARTICLE FILTER

EVALUATING SYMMETRIC INFORMATION GAP BETWEEN DYNAMICAL SYSTEMS USING PARTICLE FILTER EVALUATING SYMMETRIC INFORMATION GAP BETWEEN DYNAMICAL SYSTEMS USING PARTICLE FILTER Zhen Zhen 1, Jun Young Lee 2, and Abdus Saboor 3 1 Mingde College, Guizhou University, China zhenz2000@21cn.com 2 Department

More information

Large deviations for weighted empirical measures and processes arising in importance sampling. Pierre Nyquist

Large deviations for weighted empirical measures and processes arising in importance sampling. Pierre Nyquist Large deviations for weighted empirical measures and processes arising in importance sampling Pierre Nyquist Abstract This thesis consists of two papers related to large deviation results associated with

More information

Large deviation convergence of generated pseudo measures

Large deviation convergence of generated pseudo measures Large deviation convergence of generated pseudo measures Ljubo Nedović 1, Endre Pap 2, Nebojša M. Ralević 1, Tatjana Grbić 1 1 Faculty of Engineering, University of Novi Sad Trg Dositeja Obradovića 6,

More information

Estimating Loynes exponent

Estimating Loynes exponent Estimating Loynes exponent Ken R. Duffy Sean P. Meyn 28 th October 2009; revised 7 th January 20 Abstract Loynes distribution, which characterizes the one dimensional marginal of the stationary solution

More information

On the Complexity of Best Arm Identification with Fixed Confidence

On the Complexity of Best Arm Identification with Fixed Confidence On the Complexity of Best Arm Identification with Fixed Confidence Discrete Optimization with Noise Aurélien Garivier, Emilie Kaufmann COLT, June 23 th 2016, New York Institut de Mathématiques de Toulouse

More information

Brownian Motion. 1 Definition Brownian Motion Wiener measure... 3

Brownian Motion. 1 Definition Brownian Motion Wiener measure... 3 Brownian Motion Contents 1 Definition 2 1.1 Brownian Motion................................. 2 1.2 Wiener measure.................................. 3 2 Construction 4 2.1 Gaussian process.................................

More information

Refined Bounds on the Empirical Distribution of Good Channel Codes via Concentration Inequalities

Refined Bounds on the Empirical Distribution of Good Channel Codes via Concentration Inequalities Refined Bounds on the Empirical Distribution of Good Channel Codes via Concentration Inequalities Maxim Raginsky and Igal Sason ISIT 2013, Istanbul, Turkey Capacity-Achieving Channel Codes The set-up DMC

More information

Large Deviations for Channels with Memory

Large Deviations for Channels with Memory Large Deviations for Channels with Memory Santhosh Kumar Jean-Francois Chamberland Henry Pfister Electrical and Computer Engineering Texas A&M University September 30, 2011 1/ 1 Digital Landscape Delay

More information

How to estimate a cumulative process s rate-function

How to estimate a cumulative process s rate-function How to estimate a cumulative process s rate-function Ken Duffy (1) and Anthony P. Metcalfe (2) (1) Hamilton Institute, National University of Ireland, Maynooth, Ireland (2) Boole Centre for Research in

More information

STAT 430/510 Probability

STAT 430/510 Probability STAT 430/510 Probability Hui Nie Lecture 16 June 24th, 2009 Review Sum of Independent Normal Random Variables Sum of Independent Poisson Random Variables Sum of Independent Binomial Random Variables Conditional

More information

Statistical Modeling of Co-Channel Interference

Statistical Modeling of Co-Channel Interference Statistical Modeling of Co-Channel Interference Kapil Gulati, Aditya Chopra, Brian L. Evans and Keith R. Tinsley The University of Texas at Austin, Austin, Texas 7871 Email: gulati,chopra,bevans}@ece.utexas.edu

More information

Minimum Feedback Rates for Multi-Carrier Transmission With Correlated Frequency Selective Fading

Minimum Feedback Rates for Multi-Carrier Transmission With Correlated Frequency Selective Fading Minimum Feedback Rates for Multi-Carrier Transmission With Correlated Frequency Selective Fading Yakun Sun and Michael L. Honig Department of ECE orthwestern University Evanston, IL 60208 Abstract We consider

More information

The parabolic Anderson model on Z d with time-dependent potential: Frank s works

The parabolic Anderson model on Z d with time-dependent potential: Frank s works Weierstrass Institute for Applied Analysis and Stochastics The parabolic Anderson model on Z d with time-dependent potential: Frank s works Based on Frank s works 2006 2016 jointly with Dirk Erhard (Warwick),

More information

UCSD ECE250 Handout #27 Prof. Young-Han Kim Friday, June 8, Practice Final Examination (Winter 2017)

UCSD ECE250 Handout #27 Prof. Young-Han Kim Friday, June 8, Practice Final Examination (Winter 2017) UCSD ECE250 Handout #27 Prof. Young-Han Kim Friday, June 8, 208 Practice Final Examination (Winter 207) There are 6 problems, each problem with multiple parts. Your answer should be as clear and readable

More information

Chapter 3, 4 Random Variables ENCS Probability and Stochastic Processes. Concordia University

Chapter 3, 4 Random Variables ENCS Probability and Stochastic Processes. Concordia University Chapter 3, 4 Random Variables ENCS6161 - Probability and Stochastic Processes Concordia University ENCS6161 p.1/47 The Notion of a Random Variable A random variable X is a function that assigns a real

More information

Extreme Value Analysis and Spatial Extremes

Extreme Value Analysis and Spatial Extremes Extreme Value Analysis and Department of Statistics Purdue University 11/07/2013 Outline Motivation 1 Motivation 2 Extreme Value Theorem and 3 Bayesian Hierarchical Models Copula Models Max-stable Models

More information

Evgeny Spodarev WIAS, Berlin. Limit theorems for excursion sets of stationary random fields

Evgeny Spodarev WIAS, Berlin. Limit theorems for excursion sets of stationary random fields Evgeny Spodarev 23.01.2013 WIAS, Berlin Limit theorems for excursion sets of stationary random fields page 2 LT for excursion sets of stationary random fields Overview 23.01.2013 Overview Motivation Excursion

More information

Finite-time Ruin Probability of Renewal Model with Risky Investment and Subexponential Claims

Finite-time Ruin Probability of Renewal Model with Risky Investment and Subexponential Claims Proceedings of the World Congress on Engineering 29 Vol II WCE 29, July 1-3, 29, London, U.K. Finite-time Ruin Probability of Renewal Model with Risky Investment and Subexponential Claims Tao Jiang Abstract

More information

Thomas Knispel Leibniz Universität Hannover

Thomas Knispel Leibniz Universität Hannover Optimal long term investment under model ambiguity Optimal long term investment under model ambiguity homas Knispel Leibniz Universität Hannover knispel@stochastik.uni-hannover.de AnStAp0 Vienna, July

More information

On Consistent Hypotheses Testing

On Consistent Hypotheses Testing Mikhail Ermakov St.Petersburg E-mail: erm2512@gmail.com History is rather distinctive. Stein (1954); Hodges (1954); Hoefding and Wolfowitz (1956), Le Cam and Schwartz (1960). special original setups: Shepp,

More information

The main results about probability measures are the following two facts:

The main results about probability measures are the following two facts: Chapter 2 Probability measures The main results about probability measures are the following two facts: Theorem 2.1 (extension). If P is a (continuous) probability measure on a field F 0 then it has a

More information

Measure-Transformed Quasi Maximum Likelihood Estimation

Measure-Transformed Quasi Maximum Likelihood Estimation Measure-Transformed Quasi Maximum Likelihood Estimation 1 Koby Todros and Alfred O. Hero Abstract In this paper, we consider the problem of estimating a deterministic vector parameter when the likelihood

More information

Interactions of Information Theory and Estimation in Single- and Multi-user Communications

Interactions of Information Theory and Estimation in Single- and Multi-user Communications Interactions of Information Theory and Estimation in Single- and Multi-user Communications Dongning Guo Department of Electrical Engineering Princeton University March 8, 2004 p 1 Dongning Guo Communications

More information

On asymptotically efficient simulation of large deviation probabilities

On asymptotically efficient simulation of large deviation probabilities On asymptotically efficient simulation of large deviation probabilities A. B. Dieker and M. Mandjes CWI P.O. Box 94079 1090 GB Amsterdam, the Netherlands and University of Twente Faculty of Mathematical

More information

The Central Limit Theorem: More of the Story

The Central Limit Theorem: More of the Story The Central Limit Theorem: More of the Story Steven Janke November 2015 Steven Janke (Seminar) The Central Limit Theorem:More of the Story November 2015 1 / 33 Central Limit Theorem Theorem (Central Limit

More information

Interference and Throughput in Aloha-based Ad Hoc Networks with Isotropic Node Distribution

Interference and Throughput in Aloha-based Ad Hoc Networks with Isotropic Node Distribution Interference and Throughput in Aloha-based Ad Hoc Networks with Isotropic Node Distribution Ralph Tanbourgi, Holger Jäkel, Leonid Chaichenets and Friedrich K. Jondral Communications Engineering Lab, Karlsruhe

More information

Part IA Probability. Definitions. Based on lectures by R. Weber Notes taken by Dexter Chua. Lent 2015

Part IA Probability. Definitions. Based on lectures by R. Weber Notes taken by Dexter Chua. Lent 2015 Part IA Probability Definitions Based on lectures by R. Weber Notes taken by Dexter Chua Lent 2015 These notes are not endorsed by the lecturers, and I have modified them (often significantly) after lectures.

More information

Lecture 2: Review of Basic Probability Theory

Lecture 2: Review of Basic Probability Theory ECE 830 Fall 2010 Statistical Signal Processing instructor: R. Nowak, scribe: R. Nowak Lecture 2: Review of Basic Probability Theory Probabilistic models will be used throughout the course to represent

More information

Adaptive Monte Carlo methods

Adaptive Monte Carlo methods Adaptive Monte Carlo methods Jean-Michel Marin Projet Select, INRIA Futurs, Université Paris-Sud joint with Randal Douc (École Polytechnique), Arnaud Guillin (Université de Marseille) and Christian Robert

More information

Capacity of the Discrete Memoryless Energy Harvesting Channel with Side Information

Capacity of the Discrete Memoryless Energy Harvesting Channel with Side Information 204 IEEE International Symposium on Information Theory Capacity of the Discrete Memoryless Energy Harvesting Channel with Side Information Omur Ozel, Kaya Tutuncuoglu 2, Sennur Ulukus, and Aylin Yener

More information

Exercises with solutions (Set D)

Exercises with solutions (Set D) Exercises with solutions Set D. A fair die is rolled at the same time as a fair coin is tossed. Let A be the number on the upper surface of the die and let B describe the outcome of the coin toss, where

More information

Lecture 1: August 28

Lecture 1: August 28 36-705: Intermediate Statistics Fall 2017 Lecturer: Siva Balakrishnan Lecture 1: August 28 Our broad goal for the first few lectures is to try to understand the behaviour of sums of independent random

More information

DETECTING PHASE TRANSITION FOR GIBBS MEASURES. By Francis Comets 1 University of California, Irvine

DETECTING PHASE TRANSITION FOR GIBBS MEASURES. By Francis Comets 1 University of California, Irvine The Annals of Applied Probability 1997, Vol. 7, No. 2, 545 563 DETECTING PHASE TRANSITION FOR GIBBS MEASURES By Francis Comets 1 University of California, Irvine We propose a new empirical procedure for

More information

A Review of Basic Monte Carlo Methods

A Review of Basic Monte Carlo Methods A Review of Basic Monte Carlo Methods Julian Haft May 9, 2014 Introduction One of the most powerful techniques in statistical analysis developed in this past century is undoubtedly that of Monte Carlo

More information

Nishant Gurnani. GAN Reading Group. April 14th, / 107

Nishant Gurnani. GAN Reading Group. April 14th, / 107 Nishant Gurnani GAN Reading Group April 14th, 2017 1 / 107 Why are these Papers Important? 2 / 107 Why are these Papers Important? Recently a large number of GAN frameworks have been proposed - BGAN, LSGAN,

More information

Uncertainty. Jayakrishnan Unnikrishnan. CSL June PhD Defense ECE Department

Uncertainty. Jayakrishnan Unnikrishnan. CSL June PhD Defense ECE Department Decision-Making under Statistical Uncertainty Jayakrishnan Unnikrishnan PhD Defense ECE Department University of Illinois at Urbana-Champaign CSL 141 12 June 2010 Statistical Decision-Making Relevant in

More information

. Find E(V ) and var(v ).

. Find E(V ) and var(v ). Math 6382/6383: Probability Models and Mathematical Statistics Sample Preliminary Exam Questions 1. A person tosses a fair coin until she obtains 2 heads in a row. She then tosses a fair die the same number

More information

Semi-Parametric Importance Sampling for Rare-event probability Estimation

Semi-Parametric Importance Sampling for Rare-event probability Estimation Semi-Parametric Importance Sampling for Rare-event probability Estimation Z. I. Botev and P. L Ecuyer IMACS Seminar 2011 Borovets, Bulgaria Semi-Parametric Importance Sampling for Rare-event probability

More information

Asymptotics for posterior hazards

Asymptotics for posterior hazards Asymptotics for posterior hazards Pierpaolo De Blasi University of Turin 10th August 2007, BNR Workshop, Isaac Newton Intitute, Cambridge, UK Joint work with Giovanni Peccati (Université Paris VI) and

More information

QUANTIZATION FOR DISTRIBUTED ESTIMATION IN LARGE SCALE SENSOR NETWORKS

QUANTIZATION FOR DISTRIBUTED ESTIMATION IN LARGE SCALE SENSOR NETWORKS QUANTIZATION FOR DISTRIBUTED ESTIMATION IN LARGE SCALE SENSOR NETWORKS Parvathinathan Venkitasubramaniam, Gökhan Mergen, Lang Tong and Ananthram Swami ABSTRACT We study the problem of quantization for

More information

Trust Degree Based Beamforming for Multi-Antenna Cooperative Communication Systems

Trust Degree Based Beamforming for Multi-Antenna Cooperative Communication Systems Introduction Main Results Simulation Conclusions Trust Degree Based Beamforming for Multi-Antenna Cooperative Communication Systems Mojtaba Vaezi joint work with H. Inaltekin, W. Shin, H. V. Poor, and

More information

Lecture 5 Channel Coding over Continuous Channels

Lecture 5 Channel Coding over Continuous Channels Lecture 5 Channel Coding over Continuous Channels I-Hsiang Wang Department of Electrical Engineering National Taiwan University ihwang@ntu.edu.tw November 14, 2014 1 / 34 I-Hsiang Wang NIT Lecture 5 From

More information

δ xj β n = 1 n Theorem 1.1. The sequence {P n } satisfies a large deviation principle on M(X) with the rate function I(β) given by

δ xj β n = 1 n Theorem 1.1. The sequence {P n } satisfies a large deviation principle on M(X) with the rate function I(β) given by . Sanov s Theorem Here we consider a sequence of i.i.d. random variables with values in some complete separable metric space X with a common distribution α. Then the sample distribution β n = n maps X

More information

On the Entropy of Sums of Bernoulli Random Variables via the Chen-Stein Method

On the Entropy of Sums of Bernoulli Random Variables via the Chen-Stein Method On the Entropy of Sums of Bernoulli Random Variables via the Chen-Stein Method Igal Sason Department of Electrical Engineering Technion - Israel Institute of Technology Haifa 32000, Israel ETH, Zurich,

More information

Point Process Control

Point Process Control Point Process Control The following note is based on Chapters I, II and VII in Brémaud s book Point Processes and Queues (1981). 1 Basic Definitions Consider some probability space (Ω, F, P). A real-valued

More information

Theoretical Statistics. Lecture 1.

Theoretical Statistics. Lecture 1. 1. Organizational issues. 2. Overview. 3. Stochastic convergence. Theoretical Statistics. Lecture 1. eter Bartlett 1 Organizational Issues Lectures: Tue/Thu 11am 12:30pm, 332 Evans. eter Bartlett. bartlett@stat.

More information

Quantifying Stochastic Model Errors via Robust Optimization

Quantifying Stochastic Model Errors via Robust Optimization Quantifying Stochastic Model Errors via Robust Optimization IPAM Workshop on Uncertainty Quantification for Multiscale Stochastic Systems and Applications Jan 19, 2016 Henry Lam Industrial & Operations

More information

Ruin probabilities in multivariate risk models with periodic common shock

Ruin probabilities in multivariate risk models with periodic common shock Ruin probabilities in multivariate risk models with periodic common shock January 31, 2014 3 rd Workshop on Insurance Mathematics Universite Laval, Quebec, January 31, 2014 Ruin probabilities in multivariate

More information