To construct a perpetual self-aware cognitive agent that can continuously operate with independence, an introspective machine must be produced. To assemble such an agent, it is necessary to perform a full integration of cognition (planning, understanding, and learning) and metacognition (control and monitoring of cognition) with intelligent behaviors. I outline some key computational requirements of metacognition by describing a multi- strategy learning system called Meta-AQUA and then discuss an integration of Meta-AQUA with a nonlinear state-space planning agent. I show how the resultant system, INTRO, can independently generate its own goals, and I relate this work to the general issue of self-awareness by machine.

The American Association for Artificial Intelligence was pleased to present the AAAI 2006 Fall Symposium Series, held Friday through Sunday, October 13-15, at the Hyatt Regency Crystal City in Washington, DC. Seven symposia were held. The titles were (1) Aurally Informed Performance: Integrating Ma- chine Listening and Auditory Presentation in Robotic Systems; (2) Capturing and Using Patterns for Evidence Detection; (3) Developmental Systems; (4) Integrating Reasoning into Everyday Applications; (5) Interaction and Emergent Phenomena in Societies of Agents; (6) Semantic Web for Collaborative Knowledge Acquisition; and (7) Spacecraft Autonomy: Using AI to Expand Human Space Exploration.

We report on the eleventh annual SIGART/AAAI Doctoral Consortium, held in conjunction with the National Conference on Artificial Intelligence (AAAI-06). We discuss highlights and innovations of this year's consortium and include pointers to the consortium website.

To construct a perpetual self-aware cognitive agent that can continuously operate with independence, an introspective machine must be produced. To assemble such an agent, it is necessary to perform a full integration of cognition (planning, understanding, and learning) and metacognition (control and monitoring of cognition) with intelligent behaviors. The failure to do this completely is why similar, more limited efforts have not succeeded in the past. I outline some key computational requirements of metacognition by describing a multi- strategy learning system called Meta-AQUA and then discuss an integration of Meta-AQUA with a nonlinear state-space planning agent. I show how the resultant system, INTRO, can independently generate its own goals, and I relate this work to the general issue of self-awareness by machine.

The SNePS knowledge representation, reasoning, and acting system has several features that facilitate metacognition in SNePS-based agents. The most prominent is the fact that propositions are represented in SNePS as terms rather than as sentences, so that propositions can occur as argu- ments of propositions and other expressions without leaving first-order logic. The SNePS acting subsystem is integrated with the SNePS reasoning subsystem in such a way that: there are acts that affect what an agent believes; there are acts that specify knowledge-contingent acts and lack-of-knowledge acts; there are policies that serve as "daemons," triggering acts when certain propositions are believed or wondered about. The GLAIR agent architecture supports metacognition by specifying a location for the source of self-awareness and of a sense of situatedness in the world. Several SNePS-based agents have taken advantage of these facilities to engage in self-awareness and metacognition.

Merging operators aim at defining the beliefs/goals of a group of agents from the beliefs/goals of each member of the group. Whenever an agent of the group has preferences over the possible results of the merging process (i.e., the possible merged bases), she can try to rig the merging process by lying on her true beliefs/goals if this leads to a better merged base according to her point of view. Obviously, strategy-proof operators are highly desirable in order to guarantee equity among agents even when some of them are not sincere. In this paper, we draw the strategy-proof landscape for many merging operators from the literature, including model-based ones and formula-based ones. Both the general case and several restrictions on the merging process are considered.

We present a Bayesian framework for explaining how people reason about and predict the actions of an intentional agent, based on observing its behavior. Action-understanding is cast as a problem of inverting a probabilistic generative model, which assumes that agents tend to act rationally in order to achieve their goals given the constraints of their environment. Working in a simple sprite-world domain, we show how this model can be used to infer the goal of an agent and predict how the agent will act in novel situations or when environmental constraints change. The model provides a qualitative account of several kinds of inferences that preverbal infants have been shown to perform, and also fits quantitative predictions that adult observers make in a new experiment.

Although variants of value iteration have been proposed for finding Nash or correlated equilibria in general-sum Markov games, these variants have not been shown to be effective in general. In this paper, we demonstrate by construction that existing variants of value iteration cannot find stationary equilibrium policies in arbitrary general-sum Markov games. Instead, we propose an alternative interpretation of the output of value iteration based on a new (non-stationary) equilibrium concept that we call "cyclic equilibria." We prove that value iteration identifies cyclic equilibria in a class of games in which it fails to find stationary equilibria. We also demonstrate empirically that value iteration finds cyclic equilibria in nearly all examples drawn from a random distribution of Markov games.

We propose a mean-field approximation that dramatically reduces the computational complexity of solving stochastic dynamic games. We provide conditions that guarantee our method approximates an equilibrium as the number of agents grow. We then derive a performance bound to assess how well the approximation performs for any given number of agents. We apply our method to an important class of problems in applied microeconomics. We show with numerical experiments that we are able to greatly expand the set of economic problems that can be analyzed computationally.

We consider the scaling of the number of examples necessary to achieve good performance in distributed, cooperative, multi-agent reinforcement learning, as a function of the the number of agents n. We prove a worstcase lower bound showing that algorithms that rely solely on a global reward signal to learn policies confront a fundamental limit: They require a number of real-world examples that scales roughly linearly in the number of agents. For settings of interest with a very large number of agents, this is impractical. We demonstrate, however, that there is a class of algorithms that, by taking advantage of local reward signals in large distributed Markov Decision Processes, are able to ensure good performance with a number of samples that scales as O(log n). This makes them applicable even in settings with a very large number of agents n.