After a brief discussion of Markov Evolution Formula (MEF) expressed in Probability Bracket Notation (PBN), its close relation with the joint probability distribution (JPD) of Visible Markov Models (VMM) is demonstrated by introducing Markov State Chain Projector (MSCP). The state basis and the observed basis are defined in the Sequential Event Space (SES) of Hidden Markov Models (HMM). The JPD of HMM is derived by using basis transformation in SES. The Viterbi algorithm is revisited and applied to the famous Weather HMM example, whose node graph and inference results are displayed by using software package Elvira. In the end, the formulas of VMM, HMM and some factorial HMM (FHMM) are expressed in PBN as instances of dynamic Bayesian Networks (DBN).

Epsilon-machines are minimal, unifilar presentations of stationary stochastic processes. They were originally defined in the history machine sense, as hidden Markov models whose states are the equivalence classes of infinite pasts with the same probability distribution over futures. In analyzing synchronization, though, an alternative generator definition was given: unifilar, edge-emitting hidden Markov models with probabilistically distinct states. The key difference is that history epsilon-machines are defined by a process, whereas generator epsilon-machines define a process. We show here that these two definitions are equivalent in the finite-state case.

Reinforcement learning would enjoy better success on real-world problems if domain knowledge could be imparted to the algorithm by the modelers. Most problems have both hidden state and unknown dynamics. Partially observable Markov decision processes (POMDPs) allow for the modeling of both. Unfortunately, they do not provide a natural framework in which to specify knowledge about the domain dynamics. The designer must either admit to knowing nothing about the dynamics or completely specify the dynamics (thereby turning it into a planning problem). We propose a new framework called a partially known Markov decision process (PKMDP) which allows the designer to specify known dynamics while still leaving portions of the environment s dynamics unknown.The model represents NOT ONLY the environment dynamics but also the agents knowledge of the dynamics. We present a reinforcement learning algorithm for this model based on importance sampling. The algorithm incorporates planning based on the known dynamics and learning about the unknown dynamics. Our results clearly demonstrate the ability to add domain knowledge and the resulting benefits for learning.

A popular approach to solving a decision process with non-Markovian rewards (NMRDP) is to exploit a compact representation of the reward function to automatically translate the NMRDP into an equivalent Markov decision process (MDP) amenable to our favorite MDP solution method. The contribution of this paper is a representation of non-Markovian reward functions and a translation into MDP aimed at making the best possible use of state-based anytime algorithms as the solution method. By explicitly constructing and exploring only parts of the state space, these algorithms are able to trade computation time for policy quality, and have proven quite effective in dealing with large MDPs. Our representation extends future linear temporal logic (FLTL) to express rewards. Our translation has the effect of embedding model-checking in the solution method. It results in an MDP of the minimal size achievable without stepping outside the anytime framework, and consequently in better policies by the deadline.

We develop a qualitative theory of Markov Decision Processes (MDPs) and Partially Observable MDPs that can be used to model sequential decision making tasks when only qualitative information is available. Our approach is based upon an order-of-magnitude approximation of both probabilities and utilities, similar to epsilon-semantics. The result is a qualitative theory that has close ties with the standard maximum-expected-utility theory and is amenable to general planning techniques.

This paper investigates value function approximation in the context of zero-sum Markov games, which can be viewed as a generalization of the Markov decision process (MDP) framework to the two-agent case. We generalize error bounds from MDPs to Markov games and describe generalizations of reinforcement learning algorithms to Markov games. We present a generalization of the optimal stopping problem to a two-player simultaneous move Markov game. For this special problem, we provide stronger bounds and can guarantee convergence for LSTD and temporal difference learning with linear value function approximation. We demonstrate the viability of value function approximation for Markov games by using the Least squares policy iteration (LSPI) algorithm to learn good policies for a soccer domain and a flow control problem.

We show that maximum entropy (maxent) models can be modeled with certain kinds of HMMs, allowing us to construct maxent models with hidden variables, hidden state sequences, or other characteristics. The models can be trained using the forward-backward algorithm. While the results are primarily of theoretical interest, unifying apparently unrelated concepts, we also give experimental results for a maxent model with a hidden variable on a word disambiguation task; the model outperforms standard techniques.

Filtering---estimating the state of a partially observable Markov process from a sequence of observations---is one of the most widely studied problems in control theory, AI, and computational statistics. Exact computation of the posterior distribution is generally intractable for large discrete systems and for nonlinear continuous systems, so a good deal of effort has gone into developing robust approximation algorithms. This paper describes a simple stochastic approximation algorithm for filtering called {em decayed MCMC}. The algorithm applies Markov chain Monte Carlo sampling to the space of state trajectories using a proposal distribution that favours flips of more recent state variables. The formal analysis of the algorithm involves a generalization of standard coupling arguments for MCMC convergence. We prove that for any ergodic underlying Markov process, the convergence time of decayed MCMC with inverse-polynomial decay remains bounded as the length of the observation sequence grows. We show experimentally that decayed MCMC is at least competitive with other approximation algorithms such as particle filtering.

This presentation will introduce the audience to a new, emerging body of research on sequential Monte Carlo techniques in robotics. In recent years, particle filters have solved several hard perceptual robotic problems. Early successes were limited to low-dimensional problems, such as the problem of robot localization in environments with known maps. More recently, researchers have begun exploiting structural properties of robotic domains that have led to successful particle filter applications in spaces with as many as 100,000 dimensions. The presentation will discuss specific tricks necessary to make these techniques work in real - world domains,and also discuss open challenges for researchers IN the UAI community.

We select policies for large Markov Decision Processes (MDPs) with compact first-order representations. We find policies that generalize well as the number of objects in the domain grows, potentially without bound. Existing dynamic-programming approaches based on flat, propositional, or first-order representations either are impractical here or do not naturally scale as the number of objects grows without bound. We implement and evaluate an alternative approach that induces first-order policies using training data constructed by solving small problem instances using PGraphplan (Blum & Langford, 1999). Our policies are represented as ensembles of decision lists, using a taxonomic concept language. This approach extends the work of Martin and Geffner (2000) to stochastic domains, ensemble learning, and a wider variety of problems. Empirically, we find "good" policies for several stochastic first-order MDPs that are beyond the scope of previous approaches. We also discuss the application of this work to the relational reinforcement-learning problem.