Pierre E. Bonzon

Formal Aspects of Context

Editorial Preface D.M. Gabbay. Introduction. Formal and Computational Models of Context for Natural Language Generation K. van Deemter, J. Odijk. Requirements for Dialogue Context Modelling H. Bunt. Contextual Constraints on Thematization in Written Discourse: An Empirical Study J. Lavid. Context and Implicitness: Consequences for Traditional and Computer-assisted Text Analysis M. Galliker, D. Weimer. A Context-based Mechanization of Multi-agent Reasoning A. Cimatti, L. Serafini. Presuppositions in Context: Constructing Bridges P. Piwek, E. Krahmer. Reasoning with Multilevel Contexts in Semantic Metanetworks V.Y. Terziyan, S. Puuronen. Contextual Learning: Towards using Contexts to Achieve Generality P. Bonzon. Contextual Deontic Logic: Violation Contexts and Factual Defeasibility L.W.N. van der Torre, Y.-H. Tan. A Local Models Semantics for Propositional Attitudes F. Giunchiglia, C. Ghidini. Context-based Semantics for Information Integration L. Serafini, C. Ghidini. Structured Contexts with Fibred Semantics D.M. Gabbay, R.T. Nossum.

An Abstract Machine for Classes of Communicating Agents Based on Deduction

We consider the problem of defining executable runs for classes of communicating agents. We first define an abstract machine that generates runs for individual agents with non-deterministic plans. We then introduce agent classes whose communication primitives are based on deduction. While current communication models are overly expressive with respect to the core agent models that are used as background theory, communicating agents based on deduction achieve a balanced integration. Contrary to other more theoretical work, their operational semantics are given by an abstract machine that is defined purely in sequential terms. This machine readily offers straightforward opportunities for implementing and experimenting prototypes of collaborative agents.

Towards neuro-inspired symbolic models of cognition: linking neural dynamics to behaviors through asynchronous communications

A computational architecture modeling the relation between perception and action is proposed. Basic brain processes representing synaptic plasticity are first abstracted through asynchronous communication protocols and implemented as virtual microcircuits. These are used in turn to build mesoscale circuits embodying parallel cognitive processes. Encoding these circuits into symbolic expressions gives finally rise to neuro-inspired programs that are compiled into pseudo-code to be interpreted by a virtual machine. Quantitative evaluation measures are given by the modification of synapse weights over time. This approach is illustrated by models of simple forms of behaviors exhibiting cognition up to the third level of animal awareness. As a potential benefit, symbolic models of emergent psychological mechanisms could lead to the discovery of the learning processes involved in the development of cognition. The executable specifications of an experimental platform allowing for the reproduction of simulated experiments are given in “Appendix”.

A metacircular evaluator for a logical extension of scheme

We define a computational model for a logical extension of Scheme, and give a metacircular evaluator for it. This minimal extension incorporates two new features only, i.e. logical variables and clause expressions, which can be used to define predicates in exactly the same way as lambda expressions can be used to define functions.Higher-order properties of Scheme are preserved: predicates can be passed to and returned from function applications. Predicate applications can appear as terms in functions. On the other hand, function applications can appear as terms in predicates, and can be formal as well as actual arguments, but only as long as they can be evaluated according to the usual Scheme semantics prohibiting access to unbound variables (except for constructor applications).

Processing Functional Definitions as Declarative Knowledge: A Reduced Bytecode Implementation of a Functional Logic Machine

Given a logical extension of the Scheme language, we present a bytecode implementation of an abstract machine allowing to process functional definitions in a declarative framework. In contrast to standard Prolog machines, which include specialized instructions to unify terms, the task of building structures and lists relies on the data constructors implementing the functional part of the system. Furthermore, as the different clauses of a logical procedure are explicitely chained from within an object closure, predicate applications follow the same pattern as function applications, i.e. can be controlled by a pair of apply/return instructions. These features result in an abstract machine that is very much like a minimally extended functional machine coupled with a unification coprocessor.

Modeling Animal Brains with Evolutive Cognitive Schemas

Very specifically, functional behavior assessment is a domain in developmental psychology looking at the reasons behind a child’s observed behavior. More generally, it can be considered as the search for the explanation of human and non-human actions. Towards this goal, computational cognitive neuroscience offers a new range of possibilities that contrast with the usual statistical approaches. An attempt to assess brain functionalities in learning is illustrated here through the simulation of analogical inferences. As a main result of this paper, the mapping of evolutive cognitive schemas onto neural connection structures involving two types of cognitive transfer points out to a possible discontinuity between human and non-human minds.

Conscious Behavior through Reflexive Dialogs

We consider the problem of executing conscious behavior i.e., of driving an agent’s actions and of allowing it, at the same time, to run concurrent processes reflecting on these actions. Toward this end, we express a single agent’s plans as reflexive dialogs in a multi-agent system defined by a virtual machine. We extend this machine’s planning language by introducing two specific operators for reflexive dialogs i.e., conscious and caught for monitoring beliefs and actions, respectively. The possibility to use the same language both to drive a machine and to establish a reflexive communication within the machine itself stands as a key feature of our model.

Situated BD agents subsume BDI agents

Given an abstract architecture for agents, we propose a concrete architecture integrating a situated object and a BD deliberation. Using an operational semantics given by a logic program, it is argued that these agents subsume BDI agents.

Contextual Learning: Towards Using Contexts to Achieve Generality

A difficult task encountered in machine learning, as in many other domains, is to achieve generality. Briefly, a solution to a problem is said to be general when it is not bound to data instances describing the problem. In other words, generality allows for the abstraction of solution classes from specific problems. This article presents an attempt to use formalized contexts as a way to achieve generality in machine learning.

A Contextual Model of Beliefs for Communicating Agents

We consider the problem of reasoning about the cognitive state of communicating agents. In order to compute belief revisions resulting from information exchanges, these agents could be ascribed models of explicit mental states. Towards this end, we introduce a model of distributed beliefs based on formalizing the assertion ≪ fact P is believed by agent X in context C ≫. These developments give rise to a meta-logic program embodying the various computational aspects (object class and instance creation, proof system, message interpreter) of a complete agent test bed.