Decentralized planning in uncertain environments is a complex task generally dealt with by using a decision-theoretic approach, mainly through the framework of Decentralized Partially Observable Markov Decision Processes (DEC-POMDPs). Although DEC-POMDPS are a general and powerful modeling tool, solving them is a task with an overwhelming complexity that can be doubly exponential. In this paper, we study an alternate formulation of DEC-POMDPs relying on a sequence-form representation of policies. From this formulation, we show how to derive Mixed Integer Linear Programming (MILP) problems that, once solved, give exact optimal solutions to the DEC-POMDPs. We show that these MILPs can be derived either by using some combinatorial characteristics of the optimal solutions of the DEC-POMDPs or by using concepts borrowed from game theory. Through an experimental validation on classical test problems from the DEC-POMDP literature, we compare our approach to existing algorithms. Results show that mathematical programming outperforms dynamic programming but is less efficient than forward search, except for some particular problems. The main contributions of this work are the use of mathematical programming for DEC-POMDPs and a better understanding of DEC-POMDPs and of their solutions. Besides, we argue that our alternate representation of DEC-POMDPs could be helpful for designing novel algorithms looking for approximate solutions to DEC-POMDPs.

LEXSYS, (Legume Expert System) was a project conceived at IITA (International Institute of Tropical Agriculture) Ibadan Nigeria. It was initiated by the COMBS (Collaborative Group on Maize-Based Systems Research in the 1990. It was meant for a general framework for characterizing on-farm testing for technology design for sustainable cereal-based cropping system. LEXSYS is not a true expert system as the name would imply, but simply a user-friendly information system. This work is an attempt to give a formal representation of the existing system and then present areas where intelligent agent can be applied.

Probabilistic matrix factorization (PMF) is a powerful method for modeling data associated with pairwise relationships, finding use in collaborative filtering, computational biology, and document analysis, among other areas. In many domains, there is additional information that can assist in prediction. For example, when modeling movie ratings, we might know when the rating occurred, where the user lives, or what actors appear in the movie. It is difficult, however, to incorporate this side information into the PMF model. We propose a framework for incorporating side information by coupling together multiple PMF problems via Gaussian process priors. We replace scalar latent features with functions that vary over the space of side information. The GP priors on these functions require them to vary smoothly and share information. We successfully use this new method to predict the scores of professional basketball games, where side information about the venue and date of the game are relevant for the outcome.

Current statistical models for structured prediction make simplifying assumptions about the underlying output graph structure, such as assuming a low-order Markov chain, because exact inference becomes intractable as the tree-width of the underlying graph increases. Approximate inference algorithms, on the other hand, force one to trade off representational power with computational efficiency. In this paper, we propose two new types of probabilistic graphical models, large margin Boltzmann machines (LMBMs) and large margin sigmoid belief networks (LMSBNs), for structured prediction. LMSBNs in particular allow a very fast inference algorithm for arbitrary graph structures that runs in polynomial time with a high probability. This probability is data-distribution dependent and is maximized in learning. The new approach overcomes the representation-efficiency trade-off in previous models and allows fast structured prediction with complicated graph structures. We present results from applying a fully connected model to multi-label scene classification and demonstrate that the proposed approach can yield significant performance gains over current state-of-the-art methods.

This work describes an approach to time-series modeling of social interactions between human and robot, which is motivated by the social psychology concept of social grounding. In this model, the goal of the agents is to establish and use patterns of communication, rather than rely on existing patterns. Our goal is to allow an artifical agent to construct a pattern of shared meaning with a human or other agent through shared experience rather than relying a model provided A priori. We describe a preliminary human robot interaction study which illustrates the proposed approach.

Painters write about their what artists say about their own work, using motivations; photographers, about details of the concept mapping as a conceptual capture tool (Eskridge et.

Introducing robots as teammates in medical, space, and military domains raises interesting and challenging human factors issues that do not necessarily arise in multi-robot coordination. For example, we must consider how to design robots that integrate seamlessly with human group dynamics. An essential quality of a good human partner is her ability to robustly anticipate and adapt to other team members and the environment. Robots should preserve this ability and avoid constraining their human partners’ flexibility to act. This requires that the robot partner be capable of reasoning quickly online, and adapting to the humans’ actions in a temporally fluid way. This paper describes recent advances in dynamic plan execution, and argues that these advances provide a potentially powerful framework for explicitly modeling and efficiently reasoning on temporal information for human-robot interaction. We describe an executive named Chaski that enables a robot to coordinate with a human to execute a shared plan under different models of teamwork. We have applied Chaski to demonstrate teamwork using two Barrett Whole Arm Manipulators, and describe our ongoing work to demonstrate temporally fluid human-robot teaming using the Mobile-Dexterous-Social (MDS) robot.

Until recent years, the development of real-world humanoid robotics applications has been hampered by a lack of available mobile computational power. Unlike wheeled platforms, which can reasonably easily be expected to carry a payload of computers and batteries, humanoid robots couple a need for complex control over many degrees of freedom with a form where any significant payload complicates the balancing and control problem itself. In the last few years, however, an significant number of options for embedded processing suitable for humanoid robots have appeared (e.g. miniaturized motherboards such as beagle boards), along with ever-smaller and more powerful battery technology. Part of the drive for these embedded hardware breakthroughs has been the increasing demand by consumers for more sophisticated mobile phone applications, and these modern devices now supply much in the way of sensor technology that is also potentially of use to roboticists (e.g. accelerometers, cameras, GPS). In this paper, we explore the use of modern mobile phones as a vehicle for the sophisticated AI necessary for autonomous humanoid robots.

Information is shared more extensively when a user can confidently classify all his information according to its desired degree of disclosure prior to transmission. While high quality classification is relatively straightforward for structured data (e.g., credit card numbers, cookies, "confidential" reports), most consumer and business information is unstructured (e.g., Facebook posts, corporate email). All current technological approaches to classifying unstructured information seek to identify only that information having the desired characteristics (i.e., to maximize the percentage of filtered content that requires privacy protection). Such focus on boosting classifier Precision (P) causes technology solutions to miss sensitive information [i.e., Recall (R) is compromised for the sake of P improvement]. Such privacy protection will fall short of user expectations no matter how "intelligent" the technology may be in extending beyond keywords to user meaning. Systems must simultaneously optimize both P and R in order to protect privacy sufficiently to encourage the free flow of personal and corporate information. This requires a socio-technical methodology wherein the user is intimately involved in iterative privacy improvement. The approach is a general one in which the classifier can be modified as necessary at any time when sampling measures of P and R deem it appropriate. Matching the ever-evolving user privacy model to the technology solution (e.g., active machine learning) affords a technique for building and maintaining user trust.

Representing human behavior and cognition, from individuals to societies, presents a range of challenges to the modeling and simulation community. A common thread through many of these challenges is formulating an authentic representation of time. Many of the issues related to time representation, from the sequencing of cognitive decision processes and information processing, to communication and interaction between agents, to the longer term time scales associated with ideas such as belief revision, remain open research areas throughout the community. The inherent variability between human subjects makes generalization difficult even with data from designed experiments. Discrete event simulation (DES) provides a well-documented alternative to time-step simulation and shows potential for applications across the domain of human behavior representation. This paper provides an overview of a modular discrete event framework for social simulation, along with the social and behavioral theories underlying the currently implemented modules. We discuss the practical challenges presented by time in the representation of human cognition, and provide a case study analysis of the output of the discrete event social simulation.