La verificación estructural de sistemas basados en conocimiento.
L'aide multicritère à la décision : un cas d'intégration de critères techniques, économiques et environnementaux à Hydro-Québec
Latent Semantic Indexing using eigenvalue analysis for efficient information retrieval
Text retrieval using Latent Semantic Indexing (LSI) with truncated Singular Value Decomposition (SVD) has been intensively studied in recent years. However, the expensive complexity involved in computing truncated SVD constitutes a major drawback of the LSI method. In this paper, we demonstrate how matrix rank approximation can influence the effectiveness of information retrieval systems. Besides, we present an implementation of the LSI method based on an eigenvalue analysis for rank approximation...
Le dimostrazioni di teoremi fondate sull’uso di calcolatori
Learnheuristics: hybridizing metaheuristics with machine learning for optimization with dynamic inputs
This paper reviews the existing literature on the combination of metaheuristics with machine learning methods and then introduces the concept of learnheuristics, a novel type of hybrid algorithms. Learnheuristics can be used to solve combinatorial optimization problems with dynamic inputs (COPDIs). In these COPDIs, the problem inputs (elements either located in the objective function or in the constraints set) are not fixed in advance as usual. On the contrary, they might vary in a predictable (non-random)...
Learning Bayesian networks by Ant Colony Optimisation: searching in two different spaces.
The most common way of automatically learning Bayesian networks from data is the combination of a scoring metric, the evaluation of the fitness of any given candidate network to the data base, and a search procedure to explore the search space. Usually, the search is carried out by greedy hill-climbing algorithms, although other techniques such as genetic algorithms, have also been used.A recent metaheuristic, Ant Colony Optimisation (ACO), has been successfully applied to solve a great variety...
Learning discrete categorial grammars from structures
We define the class of discrete classical categorial grammars, similar in the spirit to the notion of reversible class of languages introduced by Angluin and Sakakibara. We show that the class of discrete classical categorial grammars is identifiable from positive structured examples. For this, we provide an original algorithm, which runs in quadratic time in the size of the examples. This work extends the previous results of Kanazawa. Indeed, in our work, several types can be associated to a word...
Learning extremal regulator implementation by a stochastic automaton and stochastic approximation theory
There exist many different approaches to the investigation of the characteristics of learning system. These approaches use different branches of mathematics and, thus, obtain different results, some of them are too complicated and others do not match the results of practical experiments. This paper presents the modelling of learning systems by means of stochastic automate, mainly one particular model of a learning extremal regulator. The proof of convergence is based on Dvoretzky's Theorem on stochastic...
Learning from imprecise examples with GA-P algorithms.
GA-P algorithms combine genetic programming and genetic algorithms to solve symbolic regression problems. In this work, we will learn a model by means of an interval GA-P procedure which can use precise or imprecise examples. This method provides us with an analytic expression that shows the dependence between input and output variables, using interval arithmetic. The method also provides us with interval estimations of the parameters on which this expression depends.The algorithm that we propose...
Learning fuzzy systems. An objective function-approach.
One of the most important aspects of fuzzy systems is that they are easily understandable and interpretable. This property, however, does not come for free but poses some essential constraints on the parameters of a fuzzy system (like the linguistic terms), which are sometimes overlooked when learning fuzzy system autornatically from data. In this paper, an objective function-based approach to learn fuzzy systems is developed, taking these constraints explicitly into account. Starting from fuzzy...
Learning imprecise semantic concepts from image databases.
In this paper we introduce a model to represent high-level semantic concepts that can be perceived in images. The concepts are learned and represented by means of a set of association rules that relate the presence of perceptual features to the fulfillment of a concept for a set of images. Since both the set of images where a perceptual feature appears and the set of images fulfilling a given concept are fuzzy, we use in fact fuzzy association rules for the learning model. The concepts so acquired...
Les problématiques de l'aide à la décision : vers l'enrichissement de la trilogie choix-tri-rangement
Les schémas mentaux : représenter et maintenir une connaissance apprise
Nous nous intéressons à l'apprentissage à partir d'exemples et à la résolution de problème dans un univers évolutif représenté par une base de connaissances incomplète. Nous formalisons un cadre de représentation de connaissances susceptible d'être élaboré et critiqué par des humains comme par des machines. Cette représentation des connaissances est appelée théorie semi-empirique car cette forme de théorie n'est pas complètement axiomatique. Nous avons formalisé la gestion de la croissance incrémentale...
Lessons learned from the implementation of a workflow management system using mobile agents.
Lexicographic combinations of preference relations in the context of Possibilistic Decision Theory.
In Possibilistic Decision Theory (PDT), decisions are ranked by a pressimistic or by an optimistic qualitative criteria. The preference relations induced by these criteria have been axiomatized by corresponding sets of rationality postulates, both à la von Neumann and Morgenstern and à la Savage. In this paper we first address a particular issue regarding the axiomatic systems of PDT à la von Neumann and Morgenstern. Namely, we show how to adapt the axiomatic systems for the pessimistic and optimistic...
Linear discriminant analysis with a generalization of the Moore-Penrose pseudoinverse
The Linear Discriminant Analysis (LDA) technique is an important and well-developed area of classification, and to date many linear (and also nonlinear) discrimination methods have been put forward. A complication in applying LDA to real data occurs when the number of features exceeds that of observations. In this case, the covariance estimates do not have full rank, and thus cannot be inverted. There are a number of ways to deal with this problem. In this paper, we propose improving LDA in this...
Linguistic knowledge base simplification regarding accuracy and interpretability.
This work proposes a new method in order to simplify linguistic knowledge bases. The main goal consists of improving simultaneously accuracy and interpretability when it is possible, or at least ensuring a good trade-off between them, as well as consistency of the final knowledge base. It is used with linguistic rules which can be defined by expert, induced by data, or both of them. The simplification process is applied to the well known wine classification problem. The results are encouraging.
L'interrogation : ébauche d'une représentation formelle par des équations
Local Boolean manifolds from knowledge representation systems.
We introduce a structure to represent observations on entities in order to obtain knowledge about some of their characteristic properties or attributes. This structure is based on the Pawlak's definition of information systems (also knowledge representation systems) and lead us to obtain algebraic structures of lattice depending from the choice of an observational context. The semantical algebraic structure so obtained is of local Boolean manifold whose global structure is an orthoposet which shows...
Local stability conditions for discrete-time cascade locally recurrent neural networks
The paper deals with a specific kind of discrete-time recurrent neural network designed with dynamic neuron models. Dynamics are reproduced within each single neuron, hence the network considered is a locally recurrent globally feedforward. A crucial problem with neural networks of the dynamic type is stability as well as stabilization in learning problems. The paper formulates local stability conditions for the analysed class of neural networks using Lyapunov's first method. Moreover, a stabilization...