P.R. Cohen and H.J. Levesque, "Confirmation and joint action," In Proceedings of the 12th International Joint Conference on Artificial Intelligence, pages 951-957, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and the robots can be heterogeneous. Therefore we will describe several team coordination frameworks which are based on the intentional cooperation model, after which their applicability in RoboCup will be discussed. 3.3. 1 Joint Intentions Theory Cohen and Levesque s joint intentions theory [13, 14, 39] is based on the notion that joint action by a group of agents is more than the union of the simultaneous individual actions, even if those actions are coordinated. It is an extension of their belief goal commitment model of mental states [12] in which intentions are specified as internal ....

P. R. Cohen and H. J. Levesque. Confirmations and joint action. In Proceedings of the 12th International Joint Conference on Artificial Intelligence, pages 951--957. Morgan Kaufman Publishers, Inc., San Mateo, California, 1991. 85

....informal framework which are used to model existing discourse from the HCRC Map Task, but we do not attempt the challenging task of generating utterances. 1. Introduction This paper investigates features of communication that may be needed to support joint activity, e.g. Clark Marshall, 1981;Cohen Levesque, 1991; Cohen Levesque, 1992; Kinny et al. 1992) We report on a close study of selected discourse segments from the HCRC Map Task Corpus (Anderson et al. 1991) in which we explored the use and perceived role of nested beliefs in this co operative, task oriented dialogue (Barrington, 1997) ....

....individual resolution with reference to a model representing particular belief nestings. The proposition can then be considered to be semantically contentful with a meaning independent of reference to a nesting of agents. Other work analysing the role of confirmations in task oriented discourse (Cohen and Levesque, 1991) highlights the need to represent a concept of joint intention. Establishing mutual belief is central to the formalism that approach adopts. Work with similarities to the framework presented here is that of Carletta with JAM building belief models relevant to discourse derived from the map task ....

P R Cohen and H Levesque. Confirmations and Joint Action. In Proceedings of the 1991 Joint Artificial Intelligence Conference, Australia, 1991, 951-957.

....rules, the smoother the interaction becomes, and some rules are clearly virtually universal. But each agent has its own set of individual rules, and we do not need to appeal to shared knowledge to account for local discourse behavior. Another common social attitude is that of Joint intention [ Cohen and Levesque, 1991 ] or shared plan [ Grosz and Sidner, 1990 ] In [ Traum, 1994 ] I also develop a similar notion, that of agents executing a multi agent plan. This is an extension of the notion of executing a plan described above. A group of agents fA i g is executing a multi agent plan MP iff: 4 1. Each A i ....

Phillip R. Cohen and Hector J. Levesque. Confirmations and joint action. In Proceedings IJCAI-91, pages 951--957, 1991.

....stands on distinguishing these concepts, e.g. Roschelle and Teasley, 1995] however others use them interchangeable, or with opposite definitions c.f. Allwood, 1976] I will not distinguish them here but will use the term collaboration without trying to engage in this ongoing debate. 2 [Cohen and Levesque, 1991a] talked about the desire for mutual belief as a motivation to perform acknowledgments, though they did not provide a model of the role the acknowledgment was supposed to play in actually achieving this mutual belief 3 Interestingly, this is based on a previous Clark model , Clark and ....

Phillip R. Cohen and Hector J. Levesque. Confirmations and joint action. In Proceedings IJCAI-91, pages 951--957, 1991.

....models of multi agent collaboration in this section. Each model will be seen to have its own unique strengths and weaknesses. These models discussed here form a baseline with which the other models in section III are compared. Joint Intentions The Joint Intentions model of Cohen and Levesque [4, 5, 6, 18] represents one of the first attempts to establish a formal theory of multi agent collaboration, and due to its clarity and expression, is probably the best known of the rational agent theories. The basic premise of their theory rests on the idea of an intention commitments (individual or joint) ....

....deliberately and intentionally, and are cognitive of their actions. Hence, we assume our agents have a PGOAL to have done an action, and believe that he was doing it while he was doing it, or (INTEND a G z) PGOAL a (DONE a BEL a (DOES a G) G) z) 1 The following discussion can be found in [6]. It is included here for completeness. 3We can then define the concept of a joint persistent goal, JPG, as 2 (JPG a b D z) MB a b D) MG a b D) MK a b (UNTIL [ MB a b D) MB a b D) MB a b z) WMG a b D) where having a mutual goal, MG, is defined by both agents believing that ....

[Article contains additional citation context not shown here]

Cohen, P. and H. Levesque. "Confirmations and Joint Action", IJCAI-91, (1991), 951-957.

....if they are not aware of the help. As we have seen before, communication adds an overhead to cooperation, but the utility function takes it into account. As a special case of helpful behavior, it is possible to predict the behaviors which are described by the definition of joint intention of [Cohen and Levesque, 1991]. In particular, we can predict that an agent, whenever he comes to know that the shared goal has been achieved, is impossible to achieve or has become irrelevant, will notify this fact to his partners. In fact, as stated above, if an agent G i knows that a partner G j has a given goal H j , G i ....

....group: in fact, communication is not predicted by our definition of cooperation. Similar reasoning can be performed to predict the notifications that an action is impossible or irrelevant. Our approach has the advantage that it does not require the assumption of perfect communication. For example, [Cohen and Levesque, 1991] set the simplificatory assumption that communication never fails, since otherwise the joint intention would be disrupted when an agent fails in notifying that he succeeded to the other partners. A group based on a shared utility function is tolerant to communication failures: if an agent who ....

Cohen, P. and Levesque, H. (1991). Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney.

....e.g. Mi passi il sale [Do you pass me the salt ] 3. Yet another possibility is to perform a question about a condition necessary for h to intend to act. One necessary condition is that he can perform the requested action; in fact, an agent adopts an intention only if he considers it feasible [6]: Int(a, do(a, act) oe cando(a, act) 3) where cando represents the joint conditions that a has the capability of doing act, believes that the applicability conditions of act hold, and the intention to do act is reconciled with his other intentions. When s asks h whether h can do an action, s ....

P.R. Cohen and H.J. Levesque. Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney, 1991.

....by the fact that it does not interfere with G j s activity. As a special case of helpful behavior, the model predicts that an agent, whenever he comes to know that the shared goal has been achieved, is impossible to achieve or has become irrelevant, will notify this fact to his partners (see [8]) In fact, as stated above, if an agent G i knows that a partner G j has a given goal H j , G i can infer that G j will also have the subsidiary goals of knowing, for example, if he succeeded. Assume that, suddenly, G i comes to know (without any further cost) that H j holds. In his next planning ....

....5) 1 (fuelA : fuelA (0 2) 1 (BelBfound : false true) 1 (more specific list of more specific actions communicateA searchA donothingA) 7 Comparison with Related Work The problem of modeling group cooperation for achieving a common goal has been widely discussed in the last decade. In [8] the authors propose a formalization of joint intention which has been the basis for much research. The main contribution of this work is the association of the notion of persistent goal with the need for the agents to coordinate their activity during the execution of a joint plan. In ....

P.R. Cohen and H.J. Levesque. Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney, 1991.

....CNR, project Pianificazione e Riconoscimento di Piani nella Comunicazione . his accordingly. Repair turns are explained as cooperative subdialogues aimed at restoring the common interpretation ground; in fact, the maintenance of a correct dialogue context is a mutual goal of the interactants [10]. Incoherent turns are not always due to misinterpretations: also topic shifts and breakdowns in cooperation should be considered. Currently, we don t model topic shifts due to the initiation of new dialogues; however, as pointed out by many researchers [11, 17, 18] focus and topic shifts are ....

....wants to satisfy a goal g; its body consists of looking for a feasible plan for g and executing it. If no feasible plan is found, the decomposition of Satisfy includes the notification to the source agent that the goal cannot be reached (see the notion of Joint Intention in agent cooperation [10]) Try execute(agt, source, action) describes the execution of actions; it includes checking the preconditions of action and performing it (possibly expanding its decomposition, if it is complex) The decomposition of Tryexecute also describes acknowledgements in agent cooperation: after ....

P.R. Cohen and H.J. Levesque. Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney, 1991.

....monitors whether the action succeeds and it replans any action whose execution fails. 2 Since the agent is embedded into a community of agents, he can ask for cooperation, as well as offer his own collaboration to his partners: therefore, we also model the collaborative behavior described in Cohen Levesque (1991), as notifying the partners about how the execution of an action proceeds. The main difference with respect to other planners is that we chose to describe the planner in a plan based declarative language. The plan based architecture allows us to apply the plan recognition algorithms developed so ....

....at achieving a speaker s goals. In this view, a sequence of speech acts is admissible when the speakers contributions aim at satisfying some goal prescribed by the team activity established previously. In turn, team activity is defined in terms of mutual beliefs about goals, and weak mutual goals (Cohen Levesque 1991): an agent involved in the collaboration should inform the partner that the goal has been achieved, is unachievable, or irrelevant, as soon as he becomes aware of this fact. Given this definition of joint intention, Smith Cohen (1996) model speech acts as attempts to form a team for pursuing ....

[Article contains additional citation context not shown here]

Cohen, P.R. & Levesque, H.J. (1991). Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney.

....focusing on the Agent Modeling plans and the procedures used to recognize and generate the agent intention driven behavior. Section 4 concludes the paper, comparing our approach to some related work. 2 The recognition of intentional behavior The discussion about intentional behavior reported in [7, 12, 20, 16] provided a theoretical background to the definition of BDI architectures, which model reactive agents with deliberative capabilities [14, 15, 27] The behavior of BDI agents is directed by an explicit representation of their beliefs, desires and intentions; however, in these architectures, the ....

....be able to abandon irrelevant intentions: this is possible in our framework because, as mentioned above, the plan instances keep the decomposition links among actions; so, if an action is given up, also the actions which are being executed as steps of its body are abandoned. As pointed out in [12], when agents cooperate to achieve their intentions, they must adopt supplementary goals in order to coordinate the agents forming the team. The second parameter of Satisfy , s (which occurs in most of the AM actions) is used to express the fact that the agent a of Satisfy is committed to the ....

[Article contains additional citation context not shown here]

P.R. Cohen and H.J. Levesque. Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney, 1991.

....to show the feasibility of our approach. The AAA uses teamwork (1) to recover a multiagent system broker failures, and (2) to maintain a specified minimum number of functional brokers in the system even when some of the brokers become inaccessible. One departure from the previous teamwork theory [7, 16] is that a team can exist independently of the identity of its members. For example, the New York Yankees are a team even if all the players are traded or it is sold to new owners. The earlier work characterized a team in terms of joint commitment between the individuals that originally ....

....mutually believes that everybody in the group has p as an individual goal (MG t p) MB t GOAL lz. member z t) p ) We now use these concepts to define team commitments that result in persistent teams. 4. Persistent teams Team activity is explained in terms of the theory of joint intentions [6, 7, 16]. This theory characterizes an agent s behavior in a team in terms of its internal state described in modal logic, linear time temporal logic, and dynamic logic of action. A joint persistent goal (JPG) expressed in terms of weak mutual goal (WMG) formalizes the notion of joint commitment. The ....

[Article contains additional citation context not shown here]

P. R. Cohen and H. J. Levesque. Confirmations and Joint Action. In Proceedings of the 12 th International Joint Conference on Artificial Intelligence, Morgan Kaufman Publishers, Inc., San Mateo, California, 1991.

No context found.

Cohen, P. R., and Levesque, H. J. 1991. Confirmations and Joint Action. In Proceedings of the 12 th International Joint Conference on Artificial Intelligence, San Mateo, Calif.

.... e and just prior to e, the actor chooses that p should eventually become true, and intends that e should produce q relative to that choice. So, p represents some ultimate goal that may or may not be achieved by the attempt, while q represents what it takes to make an honest effort (Cohen and Levesque, 1991). Definition 4. Persistent Weak Achievement Goal ( 1 2 p q) 0 1 p) 1 243 576 8 1 p) 9; 1 p) 4A B7C D 1 E FHG 1 2 p) I;J K L M 1 p) N O4P Q7R S 1 T UHV 1 2 p) W;X Y Z [ 1 q) 4 7 a 1 b cHd 1 2 q) This definition adapted from ....

Cohen, P. R., and Levesque, H. J. 1991. Confirmations and Joint Action. In Proceedings of the 12 th International Joint Conference on Artificial Intelligence, Morgan Kaufmann Publishers, Inc., San Mateo, California, pages 951-957.

....of using teamwork for fault tolerance have been presented in another paper [16] In this paper, we will be mainly concerned with the logical characterization of the recovery techniques and relating the teamwork theory to its application in the AAA. One departure from our previous theory [8, 17] is that a team can exist independently of the identity of its members. For example, the New York Yankees are a team even if all the players are traded or 1 In this paper, we use the terms broker, middle agent, and facilitator interchangeably as the work described in this paper may be extended to ....

....in section 5 and finally, we conclude in section 6 with a brief summary of the results. 2 TEAMWORK A team is more than just a collection of individuals working together in close cooperation to achieve a common goal. The agents in a team must have a shared goal as well as a shared mental state [7, 8,17]. Absence of either of these characteristics implies lack of a team irrespective of the agents having a common goal, having coordinated actions, or even their being mutually helpful. When operating normally, teams of agents may be indistinguishable from a collection of agents coordinated by ....

[Article contains additional citation context not shown here]

P. R. Cohen and H. J. Levesque. Confirmations and Joint Action. In Proceedings of the 12 th International Joint Conference on Artificial Intelligence, Morgan Kaufman Publishers, Inc., San Mateo, California, August 1991, pages 951-957.

No context found.

P. R. Cohen and H. J. Levesque, "Confirmations and joint action," in Proceedings of International Joint Conference on Artificial Intelligence, pp. 951-957, 1991.

No context found.

P.R. Cohen and H.J. Levesque, "Confirmation and joint action," In Proceedings of the 12th International Joint Conference on Artificial Intelligence, pages 951-957, 1991.

No context found.

P. R. Cohen and H. J. Levesque. Confirmation and joint action. In Procs. of IJCAI'91, pages 951--957, Sydney, 1991.

No context found.

P. R. Cohen and H. J. Levesque. Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney, 1991.

No context found.

Cohen, P. R. and Levesque, H. J., 1991. "Confirmation and Joint Action," Proceedings of International Joint Conference on Artificial Intelligence.

No context found.

Cohen, P. R. and Levesque, H. J., 1991. "Confirmation and Joint Action," Proceedings of International Joint Conference on Artificial Intelligence.

No context found.

P. R. Cohen and H. J. Levesque. Confirmation and joint action. In Procs. of IJCAI'91, pages 951--957, Sydney, 1991.

No context found.

Phillip R. Cohen and Hector J. Levesque, "Confirmations and Joint Action," In Proceedings IJCAI-91, pages 951-957, 1991.

No context found.

Cohen, P. R., and Levesque, H. J. 1991a. Confirmations and joint action. In Proceedings IJCAI-91, 951--957.

No context found.

P.R. Cohen & H.J. Levesque (1991). Confirmation and joint action. In Proc. 12th IJCAI, pages 951--957, Sydney.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC