Monte-Carlo Tree Search (MCTS) is an online planning algorithm that combines the ideas of best-first tree search and Monte-Carlo evaluation. Since MCTS is based on sampling, it does not require a transition function in explicit form, but only a generative model of the domain. Because it grows a highly selective search tree guided by its samples, it can handle huge search spaces with large branching factors. By using Monte-Carlo playouts, MCTS can take long-term rewards into account even with distant horizons. Combined with multi-armed bandit algorithms to trade off exploration and exploitation, MCTS has been shown to guarantee asymptotic convergence to the optimal policy, while providing approximations when stopped at any time. The relatively new MCTS approach has started a revolution in computer Go. Furthermore, it has achieved considerable success in domains as diverse as the games of Hex, Amazons, LOA, and Ms. Pacman; in General Game Playing, planning, and optimization. Whereas the focus of previous MCTS research has been on the practical application, current research begins to address the problem of understanding the nature, the underlying principles, of MCTS. A careful understanding of MCTS will lead to more effective search algorithms. Hence, my two interrelated research questions are: How can we formulate models that increase our understanding of how MCTS works? and How can we use the developed understanding to create effective search algorithms? This research summary describes the first steps I undertook in these directions, as well as my plans for future work.

Automatic extraction of temporal relations between event pairs is an important task for several natural language processing applications such as Question Answering, Information Extraction, and Summarization. Since most existing methods are supervised and require large corpora, which for many languages do not exist, we have concentrated our efforts to reduce the need for annotated data as much as possible. This paper presents two different algorithms towards this goal. The first algorithm is a weakly supervised machine learning approach for classification of temporal relations between events. In the first stage, the algorithm learns a general classifier from an annotated corpus. Then, inspired by the hypothesis of "one type of temporal relation per discourse'', it extracts useful information from a cluster of topically related documents. We show that by combining the global information of such a cluster with local decisions of a general classifier, a bootstrapping cross-document classifier can be built to extract temporal relations between events. Our experiments show that without any additional annotated data, the accuracy of the proposed algorithm is higher than that of several previous successful systems. The second proposed method for temporal relation extraction is based on the expectation maximization (EM) algorithm. Within EM, we used different techniques such as a greedy best-first search and integer linear programming for temporal inconsistency removal. We think that the experimental results of our EM based algorithm, as a first step toward a fully unsupervised temporal relation extraction method, is encouraging.

Existing theoretical universal algorithmic intelligence models are not practically realizable. More pragmatic approach to artificial general intelligence is based on cognitive architectures, which are, however, non-universal in sense that they can construct and use models of the environment only from Turing-incomplete model spaces. We believe that the way to the real AGI consists in bridging the gap between these two approaches. This is possible if one considers cognitive functions as a "cognitive bias" (priors and search heuristics) that should be incorporated into the models of universal algorithmic intelligence without violating their universality. Earlier reported results suiting this approach and its overall feasibility are discussed on the example of perception, planning, knowledge representation, attention, theory of mind, language, and some others.

We present a framework for a large-scale distributed eScience Artificial Intelligence search. Our approach is generic and can be used for many different problems. Unlike many other approaches, we do not require dedicated machines, homogeneous infrastructure or the ability to communicate between nodes. We give special consideration to the robustness of the framework, minimising the loss of effort even after total loss of infrastructure, and allowing easy verification of every step of the distribution process. In contrast to most eScience applications, the input data and specification of the problem is very small, being easily given in a paragraph of text. The unique challenges our framework tackles are related to the combinatorial explosion of the space that contains the possible solutions and the robustness of long-running computations. Not only is the time required to finish the computations unknown, but also the resource requirements may change during the course of the computation. We demonstrate the applicability of our framework by using it to solve a challenging and hitherto open problem in computational mathematics. The results demonstrate that our approach easily scales to computations of a size that would have been impossible to tackle in practice just a decade ago.

Since its inception, the modus operandi of multi-task learning (MTL) has been to minimize the task-wise mean of the empirical risks. We introduce a generalized loss-compositional paradigm for MTL that includes a spectrum of formulations as a subfamily. One endpoint of this spectrum is minimax MTL: a new MTL formulation that minimizes the maximum of the tasks' empirical risks. Via a certain relaxation of minimax MTL, we obtain a continuum of MTL formulations spanning minimax MTL and classical MTL. The full paradigm itself is loss-compositional, operating on the vector of empirical risks. It incorporates minimax MTL, its relaxations, and many new MTL formulations as special cases. We show theoretically that minimax MTL tends to avoid worst case outcomes on newly drawn test tasks in the learning to learn (LTL) test setting. The results of several MTL formulations on synthetic and real problems in the MTL and LTL test settings are encouraging.

The idea of sensitivity in ant colony systems has been exploited in hybrid ant-based models with promising results for many combinatorial optimization problems. Heterogeneity is induced in the ant population by endowing individual ants with a certain level of sensitivity to the pheromone trail. The variable pheromone sensitivity within the same population of ants can potentially intensify the search while in the same time inducing diversity for the exploration of the environment. The performance of sensitive ant models is investigated for solving the generalized vehicle routing problem. Numerical results and comparisons are discussed and analysed with a focus on emphasizing any particular aspects and potential benefits related to hybrid ant-based models.

Domain-independent planning is one of the foundational areas in the field of Artificial Intelligence. A description of a planning task consists of an initial world state, a goal, and a set of actions for modifying the world state. The objective is to find a sequence of actions, that is, a plan, that transforms the initial world state into a goal state. In optimal planning, we are interested in finding not just a plan, but one of the cheapest plans. A prominent approach to optimal planning these days is heuristic state-space search, guided by admissible heuristic functions. Numerous admissible heuristics have been developed, each with its own strengths and weaknesses, and it is well known that there is no single "best'' heuristic for optimal planning in general. Thus, which heuristic to choose for a given planning task is a difficult question. This difficulty can be avoided by combining several heuristics, but that requires computing numerous heuristic estimates at each state, and the tradeoff between the time spent doing so and the time saved by the combined advantages of the different heuristics might be high. We present a novel method that reduces the cost of combining admissible heuristics for optimal planning, while maintaining its benefits. Using an idealized search space model, we formulate a decision rule for choosing the best heuristic to compute at each state. We then present an active online learning approach for learning a classifier with that decision rule as the target concept, and employ the learned classifier to decide which heuristic to compute at each state. We evaluate this technique empirically, and show that it substantially outperforms the standard method for combining several heuristics via their pointwise maximum.

Game tree search algorithms such as minimax have been used with enormous success in turn-based adversarial games such as Chess or Checkers. However, such algorithms cannot be directly applied to real-time strategy (RTS) games because a number of reasons. For example, minimax assumes a turn-taking game mechanics, not present in RTS games. In this paper we present RTMM, a real-time variant of the standard minimax algorithm, and discuss its applicability in the context of RTS games. We discuss its strengths and weaknesses, and evaluate it in two real-time games.

Instead, we cluster Search is an integral part of solution approaches for NPhard training instances according to their features and determine combinatorial optimization and decision problems. Once the an assignment of branching heuristics to clusters that results ability to reason deterministically is exhausted, state-of-theart in the best performance when the branching heuristic is dynamically solvers try out different alternatives which may lead to chosen based on the current subproblem's nearest an improved (in case of optimization) or feasible (in case cluster. We test our approach on the MIP-solver Cplex that of satisfaction) solution. This consideration of alternatives we use to tackle set partitioning problems. Our experiments may take place highly opportunistically as in local search approaches, show that this approach can effectively boost search performance or systematically as in backtracking-based methods.

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.