We explore the use of end-to-end multicast traffic as measurement probes to infer network-internal characteristics. We have developed in an earlier paper [2] a Maximum Likelihood Estimator for packet loss rates on individual links based on losses observed by multicast receivers. This technique exploits the inherent correlation between such observations to infer the performance of paths between branch points in the multicast tree spanning the probe source and its receivers. We evaluate through analysis and simulation the accuracy of our estimator under a variety of network conditions. In particular, we report on the error between inferred loss rates and actual loss rates as we vary the network topology, propagation delay, packet drop policy, background traffic mix, and probe traffic type. In all but one case, estimated losses and probe losses agree to within 2 percent on average. We feel this accuracy is enough to reliably identify congested links in a wide-area internetwork. Keywords---Internet performance, end-to-end measurements, Maximum Likelihood Estimator, tomography I.

Abstract not found

This purpose of this introductory paper is threefold. First, it introduces the Monte Carlo method with emphasis on probabilistic machine learning. Second, it reviews the main building blocks of modern Markov chain Monte Carlo simulation, thereby providing and introduction to the remaining papers of this special issue. Lastly, it discusses new interesting research horizons.

We describe a simple one-person card game, patience sorting. Its analysis leads to a broad circle of ideas linking Young tableaux with the longest increasing subsequence of a random permutation via the Schensted correspondence. A recent highlight of this area is the work of Baik-Deift-Johansson which yields limiting probability laws via hard analysis of Toeplitz determinants. 1991 Mathematics Subject Classifications: Primary 60C05, 05E10, 15A52, 60F05. Research supported by N.S.F. Grant MCS 96-22859 1 Introduction This survey paper treats two themes in parallel. One theme is a purely mathematical question: describe the asymptotic law (probability distribution) of the length of the longest increasing subsequence of a random permutation. This question has been studied by a variety of increasingly technically sophisticated methods over the last 30 years. We outline three, apparently quite unrelated, methods in sections 2 - 4. The other theme is a card game, patience sorting. This gam...

This paper surveys various results about Markov chains on general (non-countable) state spaces. It begins with an introduction to Markov chain Monte Carlo (MCMC) algorithms, which provide the motivation and context for the theory which follows. Then, sufficient conditions for geometric and uniform ergodicity are presented, along with quantitative bounds on the rate of convergence to stationarity. Many of these results are proved using direct coupling constructions based on minorisation and drift conditions. Necessary and sufficient conditions for Central Limit Theorems (CLTs) are also presented, in some cases proved via the Poisson Equation or direct regeneration constructions. Finally, optimal scaling and weak convergence results for Metropolis-Hastings algorithms are discussed. None of the results presented is new, though many of the proofs are. We also describe some Open Problems.

Various notions of geometric ergodicity for Markov chains on general state spaces exist. In this paper, we review certain relations and implications among them. We then apply these results to a collection of chains commonly used in Markov chain Monte Carlo simulation algorithms, the so-called hybrid chains. We prove that under certain conditions, a hybrid chain will "inherit" the geometric ergodicity of its constituent parts. 1 Introduction A question of increasing importance in the Markov chain Monte Carlo literature (Gelfand and Smith, 1990; Smith and Roberts, 1993) is the issue of geometric ergodicity of Markov chains (Tierney, 1994, Section 3.2; Meyn and Tweedie, 1993, Chapters 15 and 16; Roberts and Tweedie, 1996). However, there are a number of different notions of the phrase "geometrically ergodic", depending on perspective (total variation distance vs. in L 2 ; with reference to a particular V -function; etc.). One goal of this paper is to review and clarify the relationship...

Abstract not found

Abstract. We consider batch reinforcement learning problems in continuous space, expected total discounted-reward Markovian Decision Problems. As opposed to previous theoretical work, we consider the case when the training data consists of a single sample path (trajectory) of some behaviour policy. In particular, we do not assume access to a generative model of the environment. The algorithm studied is policy iteration where in successive iterations the Q-functions of the intermediate policies are obtained by means of minimizing a novel Bellman-residual type error. PACstyle polynomial bounds are derived on the number of samples needed to guarantee near-optimal performance where the bound depends on the mixing rate of the trajectory, the smoothness properties of the underlying Markovian Decision Problem, the approximation power and capacity of the function set used. 1

. In this paper, we develop mathematical machinery for verifying that a broad class of general state space Markov chains reacts smoothly to certain types of perturbations in the underlying transition structure. Our main result provides conditions under which the stationary probability measure of an ergodic Harris recurrent Markov chain is differentiable in a certain strong sense. The approach is based on likelihood ratio "change-of-measure" arguments, and leads directly to a "likelihood ratio gradient estimator" that can be computed numerically. Keywords: Harris recurrent Markov chain, likelihood ratio, gradient estimation, regeneration. 1 The research of this author was supported by the U. S. Army Research Office under Contract No. DAAL03-91G -0101 and by the National Science Foundation under Contract No. DDM-9101580. 2 This author's research was supported by NSERC-Canada grant No. OGP0110050 and FCAR-Qu'ebec grant No. 93ER1654. 1. Introduction In this paper, we will study the cl...

In principle, the Bayesian approach to model selection is straightforward. Prior probability distributions are used to describe the uncertainty surrounding all unknowns. After observing the data, the posterior distribution provides a coherent post data summary of the remaining uncertainty which is relevant for model selection. However, the practical implementation of this approach often requires carefully tailored priors and novel posterior calculation methods. In this article, we illustrate some of the fundamental practical issues that arise for two different model selection problems: the variable selection problem for the linear model and the CART model selection problem.