Jerry T. Ball

The synthetic teammate project

The main objective of the Synthetic Teammate project is to develop language and task enabled synthetic agents capable of being integrated into team training simulations. To achieve this goal, the agents must be able to closely match human behavior. The initial application for the synthetic teammate research is creation of an agent able to perform the functions of a pilot for an Unmanned Aerial Vehicle (UAV) simulation as part of a three-person team. The agent, or synthetic teammate, is being developed in the ACT-R cognitive architecture. The major components include: language comprehension and generation, dialog management, agent-environment interaction, and situation assessment. Initial empirical results suggest that the agent-environment interaction is a good approximation to human behavior in the UAV environment, and we are planning further empirical tests of the synthetic teammate operating with human teammates. This paper covers the project’s modeling approach, challenges faced, progress made toward an integrated synthetic teammate, and lessons learned during development.

Evaluation of Utility of LSA for Word Sense Discrimination

The goal of the on-going project described in this paper is evaluation of the utility of Latent Semantic Analysis (LSA) for unsupervised word sense discrimination. The hypothesis is that LSA can be used to compute context vectors for ambiguous words that can be clustered together --- with each cluster corresponding to a different sense of the word. In this paper we report first experimental result on tightness, separation and purity of sense-based clusters as a function of vector space dimensionality and using different distance metrics.

A bi-polar theory of nominal and clause structure and function

It is taken as axiomatic that grammar encodes meaning. Two key dimensions of meaning that get grammatically encoded are referential meaning and relational meaning. The key claim is that, in English, these two dimensions of meaning are typically encoded in distinct grammatical poles—a referential pole and a relational pole—with a specifier functioning as the locus of the referential pole and a head functioning as the locus of the relational pole. Specifiers and heads combine to form referring expressions corresponding to the syntactic notion of a maximal projection. Lexical items and expressions functioning as modifiers are preferentially attracted to one pole or the other. If the head of an expression describes a relation, one or more complements may be associated with the head. The four grammatical functions specifier, head, modifier and complement are generally adequate to represent much of the basic structure and function of nominals and clauses. These terms are borrowed from X-Bar Theory, but they are motivated on semantic grounds having to do with their grammatical function to encode referential and relational meaning.

Toward a situation model in a cognitive architecture

The ability to coherently represent information that is situationally relevant is vitally important to perform any complex task, especially when that task involves coordinating with team members. This paper introduces an approach to dynamically represent situation information within the ACT-R cognitive architecture in the context of a synthetic teammate project. The situation model represents the synthetic teammate’s mental model of the objects, events, actions, and relationships encountered in a complex task simulation. The situation model grounds textual information from the language analysis component into knowledge usable by the agent-environment interaction component. The situation model is a key component of the synthetic teammate as it provides the primary interface between arguably distinct cognitive processes modeled within the synthetic teammate (e.g., language processing and interactions with the task environment). This work has provided some evidence that reasoning about complex situations requires more than simple mental representations and requires mental processes involving multiple steps. Additionally, the work has revealed an initial method for reasoning across the various dimensions of situations. One purpose of the research is to demonstrate that this approach to implementing a situation model provides a robust capability to handle tasks in which an agent must construct a mental model from textual information, reason about complex relationships between objects, events, and actions in its environment, and appropriately communicate with task participants using natural language. In this paper we describe an approach for modeling situationally relevant information, provide a detailed example, discuss challenges faced, and present research plans for the situation model.

Synthetic Teammate Communication and Coordination with Humans

A synthetic teammate based on ACT-R cognitive architecture has been developed to function as an Air Vehicle Operator in the context of a three-agent Unmanned Aerial Vehicle (UAV) ground control team taking part in studies in a Synthetic Task Environment (STE). In order for the synthetic teammate to function as team player with human teammates, it needs to skillfully handle the subtleties of team communication and coordination. Data from early synthetic teammate interactions with two human teammates are presented here to illustrate team communication and coordination challenges for the synthetic teammate. In turn, the synthetic teammate limitations have highlighted the intricacies involved in effective teamwork. Communication, though a terrifically challenging problem in itself, is only a foundation for coordinated teamwork or interacting as a team player.

AAAI Fall Symposium reports

The Association for the Advancement of Artificial Intelligence presented the 2007 Fall Symposium Series on Friday through Sunday, November 9–11, at the Westin Arlington Gateway, Arlington, Virginia. The titles of the seven symposia were (1) AI and Consciousness: Theoretical Foundations and Current Approaches, (2) Artificial Intelligence for Prognostics, (3) Cognitive Approaches to Natural Language Processing, (4) Computational Approaches to Representation Change during Learning and Development, (5) Emergent Agents and Socialities: Social and Organizational Aspects of Intelligence, (6) Intelligent Narrative Technologies, and (7) Regarding the “Intelligence” in Distributed Intelligent Systems.