Intelligent systems promise to amplify, augment, and extend innate human abilities. A principal example is that of assistive rehabilitation robots---artificial intelligence and machine learning enable new electromechanical systems that restore biological functions lost through injury or illness. In order for an intelligent machine to assist a human user, it must be possible for a human to communicate their intentions and preferences to their non-human counterpart. While there are a number of techniques that a human can use to direct a machine learning system, most research to date has focused on the contrasting strategies of instruction and reward. The primary contribution of our work is to demonstrate that the middle ground between instruction and reward is a fertile space for research and immediate technological progress. To support this idea, we introduce the setting of human-prompted switching, and illustrate the successful combination of switching with interactive learning using a concrete real-world example: human control of a multi-joint robot arm. We believe techniques that fall between the domains of instruction and reward are complementary to existing approaches, and will open up new lines of rapid progress for interactive human training of machine learning systems.

Members of a multi-robot team, operating within close quarters, need to avoid crashing into each other. Simple collision avoidance methods can be used to prevent such collisions, typically by computing the distance to other robots and stopping, perhaps moving away, when this distance falls below a certain threshold. While this approach may avoid disaster, it may also reduce the team's efficiency if robots halt for a long time to let others pass by or if they travel further to move around one another. This paper reports on experiments where a human operator, through a graphical user interface, watches robots perform an exploration task. The operator can manually suspend robots' movements before they crash into each other, and then resume their movements when their paths are clear. Experiment logs record the robots' states when they are paused and resumed. A behavior pattern for collision avoidance is learned, by classifying the states of the robots' environment when the human operator issues "wait" and "resume" commands. Preliminary results indicate that it is possible to learn a classifier which models these behavior patterns, and that different human operators consider different factors when making decisions about stopping and starting robots.

This paper provides a comparison between alternate styles and tecnhiques for controlling many subordinate agents: delegation vs. swarm "control" or influence. Each management style is defined and pros and cons articulated. The author then attempts to apply a model he created in prior work of the "tradeoff space" of automation control approaches along three dimensions: competence, workload and unpredictability. This application offers insights about the strengths and weaknesses of each approach, but also points to a limitation in the characterization of the tradeoff space.

Many cooperative control problems ranging from formation following, to rendezvous to flocking can be expressed as consensus problems. The ability of an operator to influence the development of consensus within a swarm therefore provides a basic test of the quality of human-swarm interaction (HSI). Two plausible approaches are : Direct- dictate a desired value to swarm members or Indirect- control or influence one or more swarm members relying on existing control laws to propagate that influence. Both approaches have been followed by HSI researchers. The Indirect case uses standard consensus methods where the operator exerts influence over a few robots and then the swarm reaches a consensus based on its intrinsic rules. The Direct method corresponds to flooding in which the operator directly sends the intention to a subset of the swarm and the command then propagates through the remainder of the swarm as a privileged message. In this paper we compare these two methods regarding their convergence time and properties in noisy and noiseless conditions with static and dynamic graphs. We have found that average consensus method (indirect control) converges much slower than flooding (direct) method but it has more noise tolerance in comparison with simple flooding algorithms. Also, we have found that the convergence time of the consensus method behaves erratically when the graph's connectivity (Fiedler value) is high.

Emotional intelligence is the key for acceptance of intelligent agents by humans as equal partners, e.g., in ad hoc teams. At the same time, its existing implementations in intelligent agents are mostly limited to basic affects. Currently, there is no consensus in the understanding of complex and social emotions at the level of functional and computational models. The approach of this work is based on the mental state formalism, originally developed as a part of the cognitive architecture GMU BICA and recently extended to include affective building blocks (A.V. Samsonovich, AAAI Technical Report WS-12-06: 109-116, 2012). In the present work, complex social emotions like humor, jealousy, compassion, shame, pride, etc. are identified as emergent patterns of appraisals represented by schemas, that capture the cognitive nature of these emotions and enable their modeling. A general model of complex emotions and emotional relationships is constructed that can be validated by simulations of emotionally biased interactions and emergent relationships in small groups of agents. The framework will be useful in cognitive architectures for designing human-like-intelligent social agents possessing a sense of humor and other human-like emotionally intelligent capabilities.

One common use for swarming agents is in social simulation. This paper reports on such a model developed to track protest activities at the May 2012 NATO summit in Chicago. The use of apoptotic stigmergic agents allows the model to run on-line, consuming two kinds of external data and reporting its results in real time.

We present an approach for learning grounded language from mixed-initiative human-robot interaction. Prior work on learning from human instruction has concentrated on acquisition of task-execution knowledge from domain-specific language. In this work, we demonstrate acquisition of linguistic, semantic, perceptual, and procedural knowledge from mixed-initiative, natural language dialog. Our approach has been instantiated in a cognitive architecture, Soar, and has been deployed on a table-top robotic arm capable of picking up small objects. A preliminary analysis verifies the ability of the robot to acquire diverse knowledge from human-robot interaction.

Learning from demonstration (LfD) is a promising technique for instructing/teaching autonomous systems based on demonstrations from people who may have little to no experience with robots. An important aspect to LfD is the communication method used to transfer knowledge from an instructor to a robot. The communication method affects the complexity of the demonstration process for instructors, the range of tasks a robot can learn, and the learning algorithm itself. We have designed a graphical interface and an instructional language to provide an intuitive teaching system. The drawback to simplifying the teaching interface is that the resulting demonstration data are less structured, adding complexity to the learning process. This additional complexity is handled through the combination of a minimal set of predefined behaviors and a task representation capable of learning probabilistic policies over a set of behaviors. The predefined behaviors consist of finite actions a robot can perform, which act as building blocks for more complex tasks.

Cooperative pathfinding is a multi-agent path planning problem where a group of vehicles searches for a corresponding set of non-conflicting space-time trajectories. Many of the practical methods for centralized solving of cooperative pathfinding problems are based on the prioritized planning strategy. However, in some domains (e.g., multi-robot teams of unmanned aerial vehicles, autonomous underwater vehicles, or unmanned ground vehicles) a decentralized approach may be more desirable than a centralized one due to communication limitations imposed by the domain and/or privacy concerns. In this paper we present an asynchronous decentralized variant of prioritized planning ADPP and its interruptible version IADPP. The algorithm exploits the inherent parallelism of distributed systems and allows for a speed up of the computation process. Unlike the synchronized planning approaches, the algorithm allows an agent to react to updates about other agents' paths immediately and invoke its local spatio-temporal path planner to find the best trajectory, as response to the other agents' choices. We provide a proof of correctness of the algorithms and experimentally evaluate them on synthetic domains.

In this paper we present the results of unstructured data clustering in this case a textual data from Reuters 21578 corpus with a new biomimetic approach using immune system. Before experimenting our immune system, we digitalized textual data by the n-grams approach. The novelty lies on hybridization of n-grams and immune systems for clustering. The experimental results show that the recommended ideas are promising and prove that this method can solve the text clustering problem.