This paper concerns learning and prediction with probabilistic models where the domain sizes of latent variables have no a priori upper-bound. Current approaches represent prior distributions over latent variables by stochastic processes such as the Dirichlet process, and rely on Monte Carlo sampling to estimate the model from data. We propose an alternative approach that searches over the domain size of latent variables, and allows arbitrary priors over the their domain sizes. We prove error bounds for expected probabilities, where the error bounds diminish with increasing search scope. The search algorithm can be truncated at any time . We empirically demonstrate the approach for topic modelling of text documents.

It has been shown recently that the complexity of belief tracking in deterministic conformant and contingent planning is exponential in a width parameter that is often bounded and small. In this work, we introduce a new width notion that applies to non-deterministic conformant and contingent problems as well. We also develop a belief tracking algorithm for non-deterministic problems that is exponential in the problem width, analyze the width of non-deterministic benchmarks, compare the new notion to the previous one over deterministic problems, and present experimental results.

One of the natural approaches for selecting actions in very From this perspective, an algorithm like RTDP fails on two large state spaces is by performing a limited amount of grounds: first, RTDP does not appear to make best use of lookahead. In the contexts of discounted MDPs, Kearns, short time windows in large state spaces; second, and more Mansour, and Ng have shown that near to optimal actions importantly, RTDP can use admissible heuristics but not informed can be selected by considering a sampled lookahead tree that base policies. On the other hand, algorithms like Policy is sufficiently sparse, whose size depends on the discount Iteration (Howard 1971), deliver all of these features except factor and the suboptimality bound but not on the number of one: they are exhaustive, and thus even to get started, problem states (Kearns, Mansour, and Ng 1999). The UCT they need vectors with the size of the state space. At the algorithm (Kocsis and Szepesvári 2006) is a version of this same time, while there are non-exhaustive versions of (asynchronous) form of Monte Carlo planning, where the lookahead trees Value Iteration such as RTDP, there are no similar are not grown depth-first but'best-first', following a selection'focused' versions of Policy Iteration ensuring anytime optimality.

General Game Playing is the design of AI systems able to understand the rules of new games and to use such descriptions to play those games effectively. Games with imperfectinformation have recently been added as a new challenge forexisting general game-playing systems. The HyperPlay technique presents a solution to this challenge by maintaining a collection of models of the true game as a foundation for reasoning, and move selection. The technique provides existing game players with a bolt-on solution to convert from perfect-information games to imperfect-information games. In this paper we describe the HyperPlay technique, show how it was adapted for use with a Monte Carlo decision making process and give experimental results for its performance.

The production of any new medicine requires solutions to many planning problems. The most fundamental of these is determining the sequence of chemical reactions necessary to physically create the drug. Surprisingly, these organic syntheses can be modeled as branching paths in a discrete, fully-observable state space, making the construction of new medicines an application of heuristic search. We describe a model of organic chemistry that is amenable to traditional AI techniques from game tree search, regression, and automatic assembly sequencing. We demonstrate the applicability of AND/OR graph search by developing the first chemistry solver to use proof-number search. Finally, we construct a benchmark suite of organic synthesis problems collected from undergraduate organic chemistry exams, and we analyze our solvers performance both on this suite and in recreating the synthetic plan for a multibillion dollar drug.

General Game Playing (GGP) agents must be capable of playing a wide variety of games skillfully. Monte-Carlo Tree Search (MCTS) has proven an effective reasoning mechanism for this challenge, as is reflected by its popularity among designers of GGP agents. Providing GGP agents with the knowledge relevant to the game at hand in real time is, however, a challenging task. In this paper we propose two enhancements for MCTS in the context of GGP, aimed at improving the effectiveness of the simulations in real time based on in-game statistical feedback. The first extension allows early termination of lengthy and uninformative simulations while the second improves the action-selection strategy when both explored and unexplored actions are available. The methods are empirically evaluated in a state-of-the-art GGP agent and shown to yield an overall significant improvement in playing strength.

The current state-of-the-art algorithm for optimal coalition structure generation is IDP-IP — an algorithm that combines IDP (a dynamic programming algorithm due to Rahwan and Jennings, AAAI'08) with IP (a tree-search algorithm due to Rahwan et al., JAIR'09). In this paper we analyse IDP-IP, highlight its limitations, and then develop a new approach for combining IDP with IP that overcomes these limitations.

A model-based diagnosis problem occurs when an observation is inconsistent with the assumption that the diagnosed system is not faulty. The task of a diagnosis engine is to compute diagnoses, which are assumptions on the health of components in the diagnosed system that explain the observation. In this paper, we extend Reiter's well-known theory of diagnosis by exploiting the duality of the relation between conflicts and diagnoses. This duality means that a diagnosis is a hitting set of conflicts, but a conflict is also a hitting set of diagnoses. We use this property to interleave the search for diagnoses and conflicts: a set of conflicts can guide the search for diagnosis, and the computed diagnoses can guide the search for more conflicts. We provide the formal basis for this dual conflict-diagnosis relation, and propose a novel diagnosis algorithm that exploits this duality. Experimental results show that the new algorithm is able to find a minimal cardinality diagnosis faster than the well-known Conflict-Directed A*.

UCT, a state-of-the art algorithm for Monte Carlo tree search (MCTS) in games and Markov decision processes, is based on UCB, a sampling policy for the Multi-armed Bandit problem (MAB) that minimizes the cumulative regret. However, search differs from MAB in that in MCTS it is usually only the final ``arm pull'' (the actual move selection) that collects a reward, rather than all ``arm pulls''. Therefore, it makes more sense to minimize the simple regret, as opposed to the cumulative regret. We begin by introducing policies for multi-armed bandits with lower finite-time and asymptotic simple regret than UCB, using it to develop a two-stage scheme (SR+CR) for MCTS which outperforms UCT empirically. Optimizing the sampling process is itself a metareasoning problem, a solution of which can use value of information (VOI) techniques. Although the theory of VOI for search exists, applying it to MCTS is non-trivial, as typical myopic assumptions fail. Lacking a complete working VOI theory for MCTS, we nevertheless propose a sampling scheme that is ``aware'' of VOI, achieving an algorithm that in empirical evaluation outperforms both UCT and the other proposed algorithms.

In the multi agent path finding problem (MAPF) paths should be found for several agents, each with a different start and goal position such that agents do not collide. Previous optimal solvers applied global A*-based searches. We present a new search algorithm called Conflict Based Search (CBS). CBS is a two-level algorithm. At the high level, a search is performed on a tree based on conflicts between agents. At the low level, a search is performed only for a single agent at a time. In many cases this reformulation enables CBS to examine fewer states than A* while still maintaining optimality. We analyze CBS and show its benefits and drawbacks. Experimental results on various problems shows a speedup of up to a full order of magnitude over previous approaches.