Catálogo de publicaciones - libros

Summary. The first part of the paper aims to stressing the analogy between conditional inference and abductive inference, making evident that in both cases what is here called “reasonable” inference involves a choice between a finite set of incompatible conclusions, selecting the most information preserving-consequent in the case of standard conditionals and the most information-preserving antecedent in the case of abductive conditionals. The consequentialist view of conditionals which is endorsed in this perspective is then extended to cover the case of higher degree conditionals, introducing in the semantical analysis the notion of inferential agents reasoning about the activity of other inferential agents. It is then shown (i) that iterated conditionals are essential in the treatment of redundant causation (ii) that abductive conditionals are essential parts of iterated conditionals in the analysis of causal preemption (iii) that there is a widespread use of second-degree conditionals involving first degree abductive conditionals. The final section is devoted to remind that Peirce’s original notion of abductive inference was actually defined in terms of second degree conditionals.

Summary. In this paper we present a conception of the Peircean pragmatic truth and a formal definition of pragmatic truth, the quasi-truth – this concept, previously introduced by da Costa and collaborators, on trying to capture the meaning of the theories of pragmatist thinkers such as Peirce and James, is considered as the truth conception inherent to empirical theories and a generalization (for partial contexts) of Tarski’s correspondence characterization of truth. By defining the mathematical concept of partial structure and by using a special semantical approach, we analyze a suitable logic that can be used as the underlying logic for theories whose truth conception is the quasi-truth. We delineate a Kripke model semantics for this logic and among some fundamental results we show that it is a kind of Ja´skowski discussive logic, a paraconsistent modal logic.

Summary. The paper presents a new control method which achieves motion control for rigid robot manipulators. It is based on sliding mode control techniques and on the compensated inverse dynamics approach. The main advantages of using sliding mode control are robustness to parameter uncertainty, insensitivity to load disturbance, and fast dynamics response, as well as a remarkable computational simplicity with respect to other robust control approaches. Furthermore the proposed approach avoids the estimation of the time-varying inertia matrix. First order and second order sliding mode control laws are presented and in both cases the problem of chattering, typical of sliding mode control, is suitably circumvented. Some simulations results are reported demonstrating the good tracking properties and performances of the proposed control strategy.

Summary. Many connections have been established in recent years between Chemistry and Computer Science, and very accurate systems, based on mathematical and physical models, have been suggested for the analysis of chemical substances. However, such a systems suffer from the difficulties of processing large amount of data, and their computational cost grows largely with the chemical and physical complexity of the investigated chemical substances. This prevent such kind of systems from their practical use in many applicative domain, where complex chemical compound are involved. In this paper we proposed a formal model, based on qualitative chemical knowledge, whose aim is to overcome such computational difficulties. The model is aimed at integrating ontological and causal knowledge about chemical compounds and compound transformations. The model allowed the design and the implementation of a system, that is based on the well known Heuristic Search paradigm, devoted to the automatically resolution of chemical formulation problems in the industrial domain of rubber compounds.

Summary. Retaining functionality of a mobile robot in the presence of faults is of particular interest in autonomous robotics. From our experiences in robotics we know that hardware is one of the weak points in mobile robots. In this paper we present the foundations of a system that automatically monitors the driving device of a mobile robot. In case of a detected fault, e.g., a broken motor, the system automatically re-configures the robot in order to allow to reach a certain position. The described system is based on a generalized model of the motion hardware. The path-planner has only to change its behavior in case of a serious damage. The high-level control system remains the same. In the paper we present the model and the foundations of the diagnosis and re-configuration system.

Summary. It is an important task to automatically extract semantic annotation of a video shot. This high level semantic information can improve the performance of video retrieval. In this paper, we propose a novel approach to annotate a new video shot automatically with a non-fixed number of concepts. The process is carried out by three steps. Firstly, the semantic importance degree (SID)is introduced and a simple method is proposed to extract the semantic candidate set (SCS) under considering SID of several concepts co-occurring in the same shot. Secondly, a semantic network is constructed using an improved K2 algorithm. Finally, the final annotation set is chosen by Bayesian inference. Experimental results show that the performance of automatically annotating a new video shot is significantly improved using our method, compared with classical classifiers such as Naïve Bayesian and K Nearest Neighbor.

Summary. Traditional approaches to computing minimal conflicts and diagnoses use search technique. It is well known that search technique may cause combination explosion. Algebraic approach may be a way to solve the problem. In this paper we present an algebraic approach to model-based diagnosis. A system with an observation can be represented by a special Petri net PN , checking whether there is a conflict between the correct system behavior and the observation corresponds to checking whether there exists a marking M ∈ R ( M 0) such that M ( p 1) and M ( p 2) are not zero, where p 1 and p 2 are labeled with the output of the system and its negation respectively. Furthermore, we show that M = M 0 + CX is such a marking, where M 0 is the initial marking, C is the incidence matrix of PN , and X is the maximal vector in {V |V is a { 0, 1 } -vector and for each transition t , if V ( t ) = 1, then there is a firing sequence t 1 , t 2 ,..., tm, t} . Then, we present an algorithm to compute the maximal vector X in V SE ( PN ) in polynomial time. Once the maximal vector in V SE ( PN ) is generated, we can check whether there is conflicts between the correct system behavior and the observation. We also present algorithms for computing minimal conflicts and diagnoses by using the above algorithm. Compared with related works, our algorithm terminates in polynomial time if the inputs of the each component in the system are not more than a given constant.

Summary. With epistemological insight and artificial intelligence techniques, our aim is to reconstruct Claude Bernard’s empirical investigations with a computational model. We suppose that Claude Bernard had in mind what we call “kernel models” that contain the basic physiological concepts upon which Claude Bernard builds his general physiological theory. The “kernel models” provide a simplified view of physiology, where the internal environment – the so-called “milieu int´erieur” –, mainly the blood, plays an essential role. According to this perspective, we assume that the “kernel models” allow Claude Bernard to make some hypotheses and to draw out their logical consequences. More precisely, the role of the “kernel models” is twofold: on the one hand, they help to generate and manage working hypotheses, for instance to enumerate the probable effects of a toxic substance, on the other hand, they derive, by simulation, the most plausible consequences of each of those hypotheses. We shall show how those “kernel models” can be specified using both description logics and multi-agent systems. Then, the paper will explain how it is possible to build, on these “kernel models”, a virtual experiment laboratory, which lets us construct and conduct virtual experiments that play a role similar to the role of thought experiments. More generally, the paper constitutes an attempt to correlate Claude Bernard’s experiments, achieved to corroborate or refute some of his working hypotheses, to virtual experiments emulated on “kernel models”.

Summary. Coltheart, Rastle, Perry, Langdon, and Ziegler [1] claim that “the psychology of reading has been revolutionized by the development of computational models of visual word recognition and reading aloud”. They attribute this to the fact that a computational model is a computer program – an algorithm – “that is capable of performing the cognitive task of interest and does so by using exactly the same information-processing procedures as are specified in a theory of how people carry out this cognitive activity” [1, p. 204]. According to this view, the computational model is the theory, not a simple instantiation of a theory. In this paper we argue that computational models of reading have indeed helped in dealing with such a complex system, in interpreting the phenomena underlying it, and in making sense of the experimental data. However, we also argue that it is crucial for a model of reading to implement a computational semantic system that is as yet a missing component of all computational models. We provide two reasons for such a move. First, this would allow explaining some phenomena arising from the interaction of semantics and lexical variables. We will review the following empirical findings: faster response times to polysemic words [2] and slower response times to synonyms [3]; the leotard [4] and turple effects [5]; and the asymmetry of the neighbourhood density effect in free and conditional reading [6]. Second, such an “enriched” model would be able to account for a richer set of tasks than current computational models do. Specifically, it would simulate tasks that require access to semantic representation to be performed, such as semantic categorization and semantically-based conditional naming. We will present a computational instantiation of a semantic module that accounts for all the described phenomena, and that has helped in generating predictions that guides on-going experimental activity.