Indeed, game theory now serves as perhaps the main analytical framework in microeconomic theory, as evidenced by its prominent role in economics textbooks (for example, Mas-Colell, Whinston, and Green 1995) and by the many Nobel prizes in economic sciences awarded to prominent game theorists. Artificial intelligence got its start shortly after game theory (McCarthy et al. 1955), and indeed pioneers such as von Neumann and Simon made early contributions to both fields (see, for example, Findler [1988], Simon [1981]). Both game theory and AI draw (nonexclusively) on decision theory (von Neumann and Morgenstern 1947); for example, one prominent view defines artificial intelligence as "the study and construction of rational agents" (Russell and Norvig 2003), and hence takes a decision-theoretic approach when the world is stochastic. However, artificial intelligence spent most of its first 40 years focused on the design and analysis of agents that act in isolation, and hence had little need for game-theoretic analysis. Starting in the mid to late 1990s, game theory became a major topic of study for computer scientists, for at least two main reasons. First, economists began to be interested in systems whose computational properties posed serious barriers to practical use, and hence reached out to computer scientists; notably, this occurred around the study of combinatorial auctions (see, for example, Cramton, Shoham, and Steinberg 2006). Second, the rise of distributed computing in general and the Internet in particular made it increasingly necessary for computer scientists to study settings in which intelligent agents reason about and interact with other agents.

We at the NASA laboratory believed that our best work came when we simultaneously advanced AI theory and provided immediately usable solutions for current NASA problems. “Solving Large-Scale Constraint Satisfaction and Scheduling Problems Using a Heuristic Repair Method,” by Steve Minton, Mark Johnston, Andy Phillips, and Phil Laird clearly achieved both. It proved that local search and repair was applicable to a wide class of constraint satisfaction problems and clearly explicated the theory behind that proof.

The Doctoral Consortium materials; a workshop for of ideas between basic and applied AI. (DC) provides an opportunity for a mentoring new faculty, instructors, IAAI-11 will consider papers in two group of Ph.D. students to discuss and and graduate students on teaching; an tracks: (1) deployed application case explore their research interests and career Educational Video Track within the studies and (2) emerging applications objectives with a panel of established AAAI-11 Video program; and a Student/Educator or methodologies.

In this article, we survey a number of mechanisms created to elicit predictions, many newly proposed within the last decade. We focus on the engineering questions: How do they work and why? What factors and goals are most important in their design? The primary goal of a prediction mechanism is to obtain and aggregate dispersed information, which often exists in tacit forms as beliefs, opinions, or judgements of agents. Coalescing information is a necessary first step for decision making in almost all domains. For example, consider seasonal influenza, a significant cause of illness and death around the world. Although it recurs every year, the geographic location, timing, magnitude, and duration of outbreaks vary widely. Many people possess relevant pieces of the full information puzzle, including doctors who meet patients, clinical microbiologists who perform respiratory culture tests, pharmacists who fill prescriptions, people who have the flu, and people who know people who have the flu.

Much of AI is concerned with the design of intelligent agents. A complementary challenge is to understand how to design “rules of encounter” by which to promote simple, robust and beneficial interactions between multiple intelligent agents. This is a natural development, as AI is increasingly used for automated decision making in real-world settings. As we extend the ideas of mechanism design from economic theory, the mechanisms (or rules) become algorithmic and many new challenges surface. Starting with a short background on mechanism design theory, the aim of this paper is to provide a nontechnical exposition of recent results on dynamic incentive mechanisms, which provide rules for the coordination of agents in sequential decision problems. The framework of dynamic mechanism design embraces coordinated decision-making both in the context of uncertainty about the world external to an agent and also in regard to the dynamics of agent preferences. In addition to tracing some recent developments, we point to ongoing research challenges.

Computational pool is a relatively recent entrant into the group of games played by computer agents. It features a unique combination of properties that distinguish it from oth- ers such games, including continuous action and state spaces, uncertainty in execution, a unique turn-taking structure, and of course an adversarial nature. This article discusses some of the work done to date, focusing on the software side of the pool-playing problem. We discuss in some depth CueCard, the program that won the 2008 computational pool tournament. Research questions and ideas spawned by work on this problem are also discussed. We close by announcing the 2011 computational pool tournament, which will take place in conjunction with the Twenty-Fifth AAAI Conference.

Software testing is an expensive process, which is vital in the industry. Construction of the test-data in software testing requires the major cost and to decide which method to use in order to generate the test data is important. This paper discusses the efficiency of search-based algorithms (preferably genetic algorithm) versus random testing, in soft- ware test-data generation. This study differs from all previous studies due to sample programs (SUTs) which are used. Since we want to in- crease the complexity of SUTs gradually, and the program generation is automatic as well, Grammatical Evolution is used to guide the program generation. SUTs are generated according to the grammar we provide, with different levels of complexity. SUTs will first undergo genetic al- gorithm and then random testing. Based on the test results, this paper recommends one method to use for automation of software testing.

When choosing a suitable technique for regression and classification with multivariate predictor variables, one is often faced with a tradeoff between interpretability and high predictive accuracy. To give a classical example, classification and regression trees are easy to understand and interpret. Tree ensembles like Random Forests provide usually more accurate predictions. Yet tree ensembles are also more difficult to analyze than single trees and are often criticized, perhaps unfairly, as `black box' predictors. Node harvest is trying to reconcile the two aims of interpretability and predictive accuracy by combining positive aspects of trees and tree ensembles. Results are very sparse and interpretable and predictive accuracy is extremely competitive, especially for low signal-to-noise data. The procedure is simple: an initial set of a few thousand nodes is generated randomly. If a new observation falls into just a single node, its prediction is the mean response of all training observation within this node, identical to a tree-like prediction. A new observation falls typically into several nodes and its prediction is then the weighted average of the mean responses across all these nodes. The only role of node harvest is to `pick' the right nodes from the initial large ensemble of nodes by choosing node weights, which amounts in the proposed algorithm to a quadratic programming problem with linear inequality constraints. The solution is sparse in the sense that only very few nodes are selected with a nonzero weight. This sparsity is not explicitly enforced. Maybe surprisingly, it is not necessary to select a tuning parameter for optimal predictive accuracy. Node harvest can handle mixed data and missing values and is shown to be simple to interpret and competitive in predictive accuracy on a variety of data sets.

The Baum-Welsh algorithm together with its derivatives and variations has been the main technique for learning Hidden Markov Models (HMM) from observational data. We present an HMM learning algorithm based on the non-negative matrix factorization (NMF) of higher order Markovian statistics that is structurally different from the Baum-Welsh and its associated approaches. The described algorithm supports estimation of the number of recurrent states of an HMM and iterates the non-negative matrix factorization (NMF) algorithm to improve the learned HMM parameters. Numerical examples are provided as well.

Markovian processes have long been used to model stochastic environments. Reinforcement learning has emerged as a framework to solve sequential planning and decision-making problems in such environments. In recent years, attempts were made to apply methods from reinforcement learning to construct decision support systems for action selection in Markovian environments. Although conventional methods in reinforcement learning have proved to be useful in problems concerning sequential decision-making, they cannot be applied in their current form to decision support systems, such as those in medical domains, as they suggest policies that are often highly prescriptive and leave little room for the user's input. Without the ability to provide flexible guidelines, it is unlikely that these methods can gain ground with users of such systems. This paper introduces the new concept of non-deterministic policies to allow more flexibility in the user's decision-making process, while constraining decisions to remain near optimal solutions. We provide two algorithms to compute non-deterministic policies in discrete domains. We study the output and running time of these method on a set of synthetic and real-world problems. In an experiment with human subjects, we show that humans assisted by hints based on non-deterministic policies outperform both human-only and computer-only agents in a web navigation task.