Gmytrasiewicz, P. J., and Durfee, E. H. 1993. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation 2:237--258.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the distance to the goal, resulting in less search but a possibly non optimal solution. If an agent intentionally misstates its boundaries, whether its motives are benevolent or malicious, issues of deception arise. Deception in multiagent environments is an important and ongoing area of research [5, 10, 28] that must be addressed in the larger scope of multiagent systems in general; however, in the STEAM system, agents are assumed to never lie. In addition to the constraint form clause of a constraint, the boundary constraints used by the STEAM agents include a flexibility attribute, ....

P.J. Gmytrasiewicz and E.H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 2(3):237--258, 1993.

....payoff from 3.5 to 4. Moreover, if it is the case that R1 has correct model(s) of R2 s payoffs and R2 s rationality, then if R1 can transmit an intention that it can rationally live up to (as above) the two agents must be in a Nash equilibrium (assuming the message is received and believed (Gmytrasiewicz and Durfee, 1993)) Theorem 2: In the two agent case, if the speaker has correct model(s) of the payoffs and rationality of the hearer , then if the speaker transmits an intention that is rational in the context of how the hearer will respond to that intention, then the two agents must be in equilibrium. Proof ....

Gmytrasiewicz, P.J., and Durfee, E.H. 1993. Toward a Theory of Honesty and Trust Among Communicating Autonomous Agents. Group Decision and Negotiation 2:237-258.

....each of these approaches has its merits, each also involves some cost or risk: e#ective communication requires careful decision making about what to say, when to say it, and whether to trust what is heard (Durfee, # Supported, in part, by NSF grant IRI 9158473. Gmytrasiewicz, Rosenschein 1994) (Gmytrasiewicz Durfee 1993); relying on learned patterns of action (Sen Durfee 1994) risks jumping to incorrect expectations when environmental variations occur; and using deeper models of other agents can be more accurate but extremely time consuming (Gmytrasiewicz 1992) In this paper, we concentrate on coordinated ....

....and directions for further work, discussed in the Conclusion. RMM and the Pursuit Task: The basic mod1 eling primitives we use are based on the Recursive Modeling Method (RMM) Gmytrasiewicz 1992; Gmytrasiewicz, Durfee, Wehe 1991; Durfee, Gmytrasiewicz, Rosenschein 1994; Durfee, Lee, Gmytrasiewicz 1993). RMM provides a theoretical framework for representing and using the knowledge that an agent has about its expected payo#s and those of others. To use RMM, an agent is expected to have a payo# matrix where each entry represents the payo#s the agent expects to get given the combination of actions ....

Gmytrasiewicz, P. J., and Durfee, E. H. 1993. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation 2:237--258.

....of other agents. They are rigorously defined in [16, 17] For the purpose of the present discussion we assume that the truthfulness of these messages is guaranteed (as we mentioned the agents intentions may change, but it is in their best interest to inform others of such changes. See also [15] for cases involving lying) Thus, a hearer can use an intentional message to predict what the speaker will do. In modeling the hearer, therefore, the speaker can truncate the projected recursive structure, because it knows that the hearer s conjecture of the speaker s actions correspond exactly ....

Piotr J. Gmytrasiewicz and Edmund H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 2:237--258, 1993.

.... to confirm the role and importance of mental models, including nested models, of other agents, and how the ability to form and process these models sets humans apart from other primates 1 , and we used them in our previ 1 For example, adult humans can reliably (with 5 15 error rates) ous work [7, 8, 9, 10, 20]. Given that the ability to communicate can be advantageous, the agents may want to enrich their communicative capabilities. Specifically, if it happens that two interacting agents do not share a common agent communication language (ACL) they may want to initiate its creation and enrichment to ....

P. J. Gmytrasiewicz and E. H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 2:237--258, 1993.

....other agents about it. Definition 5: Intentional communicative acts contain information about the intentional probabilities p (R i ;ff) R j a k p defined in the companion paper. For the purpose of the present discussion we assume that the truthfulness of these messages is guaranteed (see [9] for cases involving lying) Thus, a hearer can use this message to predict exactly what the speaker will do. In modeling the hearer, therefore, the speaker can truncate the projected recursive structure, because it knows that the hearer s conjecture of the speaker s actions correspond exactly to ....

Piotr J. Gmytrasiewicz and Edmund H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 2:237--258, 1993.

....transformations that the messages induce on the state of the agent s knowledge state and the solution of multiple recursive hierarchies resulting from these transformations. The effect of relaxing the assumptions about the messages being truthful and always believed adds even more overhead (see [7]) The extension we presented in this paper makes considerable computational demands as well. The knowledge oriented actions we are considering could be embedded in dialogue trees of essentially unlimited size, and, therefore, using the full blown definition of their expected utility might require ....

Piotr J. Gmytrasiewicz and Edmund H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 1993, to appear.

....dimension in the communicative behavior among autonomous agents arises when the assumptions (implicit so far) that the messages are always true and always believed, are relaxed. We have made a preliminary investigation of these issues, and some of our most encouraging results are presented in [7], but we do not elaborate on them here. 5 Conclusion We have presented a personal, biased view of some of the main philosophical currents in reasoning about other agents, the relevant theories, and our preliminary implementation, with the aim of illustrating how these diverse threads can be ....

Piotr J. Gmytrasiewicz and Edmund H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 1993, to appear.

....consider what would happen in our original scenario in Figure 1 if R 1 were to send a message M 1 stating There is an observation point P2, twice as high as P1, behind the trees . If we assume that communication channels never lose messages (but see [10] and messages are always believed (see [7], then R 1 can be sure that R 2 will know about the point P2 as a result of the message having been sent. Thus, the recursive model structure will change due to the pragmatic meaning of M 1 , as depicted in Figure 3. Before the message was sent, R 1 s best option was a 1 2 , that is, to ....

....is what it would have done anyway) otherwise. 4 4 Of course, if R1 were to value sending R2 on a wild goose chase, and R2 did not know about this propensity, then R1 could successfully lie to R2 . For further investigation into honesty and trust among rational agents, the reader is referred to [7, 19, 20]. Imperatives as Mitigating the Costs of Computation To understand imperatives requires a similar analysis, except that now we also have to take into account the costs (time, effort) of decision making. In a nutshell, it is rational for an agent to obey an order from another if the default ....

[Article contains additional citation context not shown here]

Piotr J. Gmytrasiewicz and Edmund H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 2:237--258, 1993.

....patterns of actions taken by others, or on deeper models of other agents state. While each of these approaches has its merits, each also involves some cost or risk: effective communication requires careful decision making about what to say, when to say it, and whether to trust what is heard [2] [9]; relying on learned patterns of action [15] risks jumping to incorrect expectations when environmental variations occur; and using deeper models of other agents can be more accurate but extremely time consuming [7] In this paper, we concentrate on coordinated decision making using deeper, nested ....

Piotr J. Gmytrasiewicz and Edmund H. Durfee. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation, 2:237--258, 1993.

....of these approaches has its merits, each also involves some cost or risk: effective communication requires careful decision making about what to say, when to say it, and whether to trust what is heard (Durfee, Supported, in part, by NSF grant IRI 9158473. Gmytrasiewicz, Rosenschein 1994) (Gmytrasiewicz Durfee 1993); relying on learned patterns of action (Sen Durfee 1994) risks jumping to incorrect expectations when environmental variations occur; and using deeper models of other agents can be more accurate but extremely time consuming (Gmytrasiewicz 1992) In this paper, we concentrate on coordinated ....

....and directions for further work, discussed in the Conclusion. RMM and the Pursuit Task: The basic mod eling primitives we use are based on the Recursive Modeling Method (RMM) Gmytrasiewicz 1992; Gmytrasiewicz, Durfee, Wehe 1991; Durfee, Gmytrasiewicz, Rosenschein 1994; Durfee, Lee, Gmytrasiewicz 1993). RMM provides a theoretical framework for representing and using the knowledge that an agent has about its expected payoffs and those of others. To use RMM, an agent is expected to have a payoff matrix where each entry represents the payoffs the agent expects to get given the combination of ....

Gmytrasiewicz, P. J., and Durfee, E. H. 1993. Toward a theory of honesty and trust among communicating autonomous agents. Group Decision and Negotiation 2:237--258.

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