On need-driven proactive information exchanges in agent teams (original) (raw)

Proactive information exchanges based on the awareness of teammates��� information needs

2003

The capabilities for agents in a team to anticipate information needs of teammates and proactively offer relevant information are highly desirable. However, such behaviors have not been fully prescribed by existing agent theories. We attempt to establish a theory about proactive information exchanges based on the SharedPlan framework and Cohen and Levesque's formalization of communicative actions. We first formally specify two types of information needs. A new performative called P roInf orm is introduced by extending the semantics of Inf orm to include the speaker's belief about the information needs of the addressee. For agents in a team containing subteams to achieve proactive information exchanges, we define the semantics of "subscribe" through the third party (e.g., a broker agent). We also show that proactive information exchanges using these communicative actions can be derived as assist behaviors from the theory. The framework not only serves as a formal specification for designing agent architectures, algorithms, and applications that support proactive information exchanges among agents in a team, but also offers opportunities for extending existing agent communication protocols to support proactive teamwork.

A theoretical framework on proactive information exchange in agent teamwork

Artificial Intelligence, 2005

Proactive information delivery is critical to achieving effective teamwork. However, existing theories do not adequately address proactive information delivery. This paper presents a formal framework for proactive information delivery in agent teamwork. First, the concept of information need is introduced. Second, a new modal operator, InfoNeed is used to represent information needs. The properties of the InfoNeed operator and its relationships to other mental modal operators are examined, four types of information needs are formally identified, and axioms for anticipating the information needs of other agents are proposed and justified. Third, the axiom characterizing chains of helpful behavior in large agent teams is given. Fourth, the semantics for two proactive communicative acts (ProInform and 3PTSubscribe) is given using a reformulation of the Cohen-Levesque semantics for communicative acts in terms of the SharedPlans formalism of Grosz and Kraus. The work in this paper not only provides a better understanding of the underlying assumptions required to justify proactive information delivery behavior, but also provides a coherent basis for the specification and design of agent teams with proactive information delivery capabilities.

On proactive delivery of needed information to teammates

2002

Psychological studies about human teamwork have shown that members of an effective team can often anticipate needs of other teammates and take appropriate actions accordingly. CAST is a teamwork model that enables agents in a team to anticipate information needs of teammates, whether they are software agents or human agents. Based on such needs, agents can choose to assist teammates through proactive communications and information delivery. In this paper, we establish the formal foundation of such proactive behavior using SharedPlan theory. We show that the proactive information delivery behavior of agents can be derived from the assist axiom in SharedPlan theory. This formal foundation of proactive information delivery behavior is critical not only for understanding the underlying assumptions required to justify the behavior but also for studying the impact of an agent's belief about other teammates' observability on the agent's choice for proactive information delivery actions.

Information needs in agent teamwork

2004

Members of effective human teams can often anticipate information needs of teammates and offer relevant information to them proactively. Such capabilities are highly desirable for agent teams to achieve better teamwork processes for supporting information gathering, information fusion, and decision makings of teammates. However, there is a lack of agent theories for specifying such proactive agent behavior. The starting point of establishing such a theory is to formally characterize the concept of "information-need" and provide a framework for reasoning about others' information-needs. To this end, in this paper we (1) introduce a modal operator to represent agents' information-needs; (2) investigate levels of information-needs using the idea of precondition-tree; (3) identify several types of information-needs prevalent in agent teamwork; (4) provide and justify the axioms for anticipating others' information-needs; and (5) to complete the framework, introduce an axiom for enabling agents to commit to helping others with their information-needs. This paper thus provides a formal basis for developing agent theories about proactive information delivery behavior.

Proactive Communications in Agent Teamwork

Lecture Notes in Computer Science, 2004

The capabilities for agents in a team to anticipate informationneeds of teammates and proactively offer relevant information are highly desirable. However, such behaviors have not been fully prescribed by existing agent theories. To establish a theory about proactive information exchanges, we first introduces the concept of "information-needs", then identify and formally define the intentional semantics of two proactive communicative acts, which highly depend on the speaker's awareness of others' information-needs. It is shown that communications using these proactive performatives can be derived as helping behaviors. Conversation policies involving these proactive performatives are also discussed. The work in this paper may serve as a guide for the specification and design of agent architectures, algorithms, and applications that support proactive communications in agent teamwork.

On Modeling and Simulating Agent Teamwork in Cast

Active Media Technology - Proceedings of the Second International Conference, 2003

Effective human teams use overlapping shared mental models for anticipating information needs of teammates and for offering relevant information proactively. The long-term goal of our research is to empower agents with such "shared mental models" so that they can be used to better simulate, train, or support human teams for their information fusion, interpretation, and decisions. Toward this goal, we have developed a team agent architecture called CAST that enables agents to infer information needs of teammates, which further enables agents to assist teammates by proactively delivering needed information to them. In this paper, we focus on two key issues related to proactive information delivery behavior. First, we model the semantics of proactive information delivery as an attempt (called ProAssert), which extends the performative Assert in Joint Intention Theory. Second, we introduce a decision-theoretic approach for reasoning about whether to act on a potential proactive assert. Experimental results suggested that the decision-theoretic communication strategy enhances the team performance. The formal semantics and the decision-theoretic communication strategies together provide a sound and practical framework that enables further studies regarding proactive information delivery for supporting the decision making of a team involving human and agents.

Towards Team-Orientation in Agent Design: Social Plan Execution

2003

We present a process-algebraic approach for the specification of agent systems where agents participate in joint activities, extending previous work by the first author in . While related to existing work on teamwork, such as , our focus here is not on discussing notions bearing on joint intentions or abilities. Rather, we focus on providing a specification language for agent systems and behaviors of agents. The language of behaviors we propose can also serve as a coordination language for specifying the flow of control in joint agent activity. We then provide an operational semantics for the language of social agents that we present, based on a notion of social structure that simplifies the semantics of social plan execution. The formation of the required social structure is part of the process of social plan formation. In this report we assume social plans are already available (pre-compiled or already formed) and focus on providing semantics for their execution.

A decision-theoretic approach for designing proactive communication in multi-agent teamwork

Proceedings of the 2004 ACM symposium on Applied computing - SAC '04, 2004

Techniques that support effective communication during teamwork processes are of particular importance. Psychological study shows that an effective team often can anticipate information exchange among the team and communicate relevant information proactively. Proactive communication is crucial for understanding and sharing common goals and for cooperative actions. Communication can be valuable if it assists agents with new and timely information; it also has cost because it consumes network resources such as bandwidth. To address these issues, we present a new model that uses information production and need to capture the complex multi-agent communication process and a dynamic decision-theoretic determination of communication strategies. We also introduce a generic utility function and an algorithm, DTPC (Decision-Theoretic Proactive Communication), that focuses on representing information production and need of team members and resolving decision interactions among them for making decisions.

A Framework for Agent Collaboration in

2002

In this research, we introduce the development of a framework that supports collaboration between a group of heterogeneous agents, supporting the development and maintenance of mutual understanding and a shared view of the task being solved. The developed framework builds on two well known formal models of teamwork namely, the Joint Intentions Theory proposed by P. Cohen and H. Levesque and the Cooperative Problem Solving Model by M. Wooldridge and N. Jennings, and builds on layered agent conversational model defined by M. Nowostawski, M. Purvis and S. Cranefield. Rather than assuming specific agent architecture, the developed collaboration framework develops a set of specifications that an agent should follow in order to guarantee its success as a team member. In this sense, the proposed approach focuses on creating a high-level collaboration protocol for developing and maintaining a mutual understanding and a shared view of teamwork. In addition, the developed framework is transparent to agent architecture, organization and interaction protocols, enabling it to be used in different domains and on several platforms, and allowing integration with various agent development environments, and allowing interoperability between agents having different agent architectures. The developed framework maintains a clear separation between the conceptual model, defined once, and the possible model implementation, where implementations introduce details of messaging, ontology, communication and interaction. In this research, we propose and implement a framework that we utilize in the development and implementation of the case study, which is the problem of performing trade exchange and settlement between a team of trading agents.

Proactive information exchange during team cooperation

2002

We are concerned, ultimately, with multi-agent teams involving both software and human agents in which the software agents assist in training the humans, e.g., as virtual team members. The software agents must be capable of utilizing the teamwork mechanisms used by humans. In this paper, we describe how to give agents the same implicit communication capabilities, i.e., observation, that humans use. We show how agents can use observations of the environment and teammates' actions to estimate other agents' beliefs without bothering them with unnecessary messages; we also show how agents can anticipate information needs among the team and proactively exchange information, reducing the total communication volume. To achieve these goals, we add sensing capabilities to agents and present algorithms that infer teammates' mental state from the observation. Experimental data is presented that shows that the advantage of observation in proactive information exchange as well as enhanced team performance.