Benoit Gaudou

A Logical Framework for Grounding-based Dialogue Analysis

A major critique against BDI (Belief, Desire, Intention) approaches to communication is that they require strong hypotheses such as sincerity and cooperation on the mental states of the agents (cf. for example [Singh, M.P., Agent Communication Languages: Rethinking the Principles, Computer 31 (1998), pp. 40-47; Singh, M.P., A Social Semantics for Agent Communication Languages, in: F. Dignum and M. Greaves, editors, Issues in Agent Communication, number 1916 in LNAI (2000), pp. 31-45; Fornara, N. and M. Colombetti, Operational Specification of a Commitment-Based Agent Communication Language, in: C. Castelfranchi and L. W. Johnson, editors, Proc. First Int. Joint Conf. on Autonomous Agents and MultiAgent Systems (AAMAS-2002), ACM Press 2, 2002, pp. 535-542]). The aim of this paper is to remed this defect. Thus we study communication between heterogeneous agents via the notion of grounding, in the sense of being publicly expressed and established. We show that this notion is different from social commitment, from the standard mental attitudes, and from different versions of common belief. Our notion is founded on speech act theory, and it is directly related to the expression of the sincerity condition [Searle, J.R., Speech acts: An essay in the philosophy of language, Cambridge University Press, New York, 1969; Searle, J.R., Intentionality: An essay in the philosophy of mind, Cambridge University Press, 1983; Vanderveken, D., Principles of language use, Meaning and Speech Acts 1, Cambridge University Press, 1990] when a speech act is performed. We use this notion to characterize speech acts in terms of preconditions and effects. As an example we show how persuasion dialogues a la Walton & Krabbe can be analyzed in our framework. In particular we show how speech act preconditions constrain the possible sequences of speech acts.

OCC’s emotions: a formalization in a BDI logic

Nowadays, more and more artificial agents integrate emotional abilities, for different purposes: expressivity, adaptability, believability... Designers mainly use Ortony et al.’s typology of emotions, that provides a formalization of twenty-two emotions based on psychological theories. But most of them restrain their agents to a few emotions among these twenty-two ones, and are more or less faithful to their definition. In this paper we propose to extend standard BDI (belief, desire, intention) logics to account for more emotions while trying to respect their definitions as exactly as possible.

Agents BDI et simulations sociales, Unis pour le meilleur et pour le pire

La modelisation et simulation ont longtemps ete dominees par les approches basees sur les equations, jusqua lavenement recent des approches orientees agents. Pour freiner laugmentation de complexite des modeles que peut entrainer lutilisation de cette nouvelle approche, la tendance est a la sursimplification des modeles. Des modeles plus descriptifs ont cependant ete developpes pour une variete de phenomenes, mais la cognition des agents est encore trop souvent negligee alors quelle a une grande importance dans certains domaines, en particulier en sciences humaines et sociales. La solution que nous proposons dans cet article est dutiliser des agents BDI. Nous montrons quil sagit dun paradigme expressif, realiste et simple qui apporte de nombreux benefices a la simulation a base dagents.

A cluster-based approach for disturbed, spatialized, distributed information gathering systems

In a distributed spatialized information collecting system managed by a swarm of agents, where some are supposed disturbed, the maintenance of the system coherence and cooperation between reliable elements is a challenge. This paper tackles the problem of finding an efficient mechanism to ensure the coherence of the system and to optimize system performance. The main contribution of this paper consists of two major steps: (i) use trust-based mechanism to ensure the coherence and the robustness of the system; (ii) allow reliable elements to create dynamic clusters based on trust. We propose two different organizations in order to manage these issues and show how they must interact: a social one in which each agent maintains a TrustSet to estimate trust on others; a spacial one in which reliable elements are grouped in an ad hoc type network to improve cooperation between themselves.

Anchoring the Institutional Dimension of Speech Acts in Agents' Attitudes: A Logical Approach

The focus of this contribution is to define speech acts as institutional actions where modeled institutions are social or informal institutions. With social or informal institution, we mean an institution which is grounded on the acceptances of its members. In the first part of the paper we present a logic of acceptance, goal and action. Then, we specify how agents can create and maintain normative and institutional facts on the basis of their acceptances qua members of a certain institution. In particular we focus on obligations and social commitments and provide an original reductionist characterization of these concepts, anchoring them in agents attitudes. Finally, we propose a formal characterization of the speech act promise. In this last part, we first define the constitutive rule which creates the relation between an utterance (as a physical action) and the speech act promise (as an institutional action). Then, we specify the deontic dimension of promise. Finally, we establish the relationship between promise and social commitment.