Supervised Machine Learning Applied to Link Prediction in Bipartite Social Networks.

A supervised machine learning link prediction approach for academic collaboration recommendation.

Recherche adaptative de structures de r%E9gulation g%E9n%E9tique.

LICORN learning cooperative regulation networks from gene expression data.

Fouille de donn%E9es pour l%27extraction de grands r%E9seaux de r%E9gulation g%E9n%E9tique.

Fouille de donn%E9es pour l%27extraction de grands r%E9seaux de r%E9gulation g%E9n%E9tique.

Extension of the Top%2DDown Data%2DDriven Strategy to ILP.

VAMP Visualization and analysis of array%2DCGH
Assessing variations in DNA copy number is crucial for understanding constitutional or somatic diseases, particularly cancers. The recently developed array-CGH (comparative genomic hybridization) technology allows this to be investigated at the genomic level. We report the availability of a web tool for analysing array-CGH data. CAPweb (CGH array Analysis Platform on the Web) is intended as a user-friendly tool enabling biologists to completely analyse CGH arrays from the raw data to the visualization and biological interpretation. The user typically performs the following bioinformatics steps of a CGH array project within CAPweb: the secure upload of the results of CGH array image analysis and of the array annotation (genomic position of the probes); first level analysis of each array, including automatic normalization of the data (for correcting experimental biases), breakpoint detection and status assignment (gain, loss or normal); validation or deletion of the analysis based on a summary report and quality criteria; visualization and biological analysis of the genomic profiles and results through a user-friendly interface. CAPweb is accessible at
Computation of recurrent minimal genomic alterations from array%2DCGH data.

Computation of recurrent minimal genomic alterations from array%2DCGH data.

Apprentissage de signatures de facteurs de transcription %E0 partir de donn%E9es d%27expression.

Local Pattern Discovery in Array%2DCGH Data.

Constraint%2Dbased Learning of Long Relational Concepts.

Inductive Logic Programming
 
Lazy Propositionalisation for Relational Learning.
A number of Inductive Logic Programming (ILP) systems have addressed the problem of learning First Order Logic (FOL) discriminant definitions by first reformulating the FOL learning problem into an attribute-value one and then applying efficient learning techniques dedicated to this simpler formalism. The complexity of such propositionalisation methods is now in the size of the reformulated problem which is exponential when tackling non-determinate relational problems. We propose a method that selectively propositionalises the FOL training set by interleaving attribute-value reformulation and algebraic resolution. It avoids, as much as possible, the generation of reformulated examples which are not relevant wrt the discrimination task, and still ensures that explicit correct and complete definitions are learned. We present an AQ-like algorithm exploiting this lazy propositionalisation method and then provide a first empirical evaluation on a standard benchmark dataset for ILP, the Mutagenesis problem. 1
Towards Learning in CARIN%2DALN.
. In this paper we investigate a new language for learning,  which combines two well-known representation formalisms, Description  Logics and Horn Clause Logics. Our goal is to study the feasability of  learning in such a hybrid description - horn clause language, namely  CARIN-ALN [LR98b], in the presence of hybrid background knowledge,  including a Horn clause and a terminological component. After setting  our learning framework, we present algorithms for testing example coverage  and subsumption between two hypotheses, based on the existential  entailment algorithm studied in[LR98b]. While the hybrid language is  more expressive than horn clause logics alone, the complexity of these  two steps for CARIN-ALN remains bounded by their respective complexity  in horn clause logics.  1 
Any%2Dtime Relational Reasoning Resource%2Dbounded Induction and Deduction Through Stochastic Matching.

Selective Propositionalization for Relational Learning.

Machine Learning ECML%2D98

Natural Ideal Operators in Inductive Logic Programming.

Natural Ideal Operators in Inductive Logic Programming.

Tractable Induction and Classification in First Order Logic Via Stochastic Matching.
Learning in first-order logic (FOL) languages suffers from a specific difficulty: both induction and classification are potentially exponential in the size of hypotheses. This difficulty is usually dealt with by limiting the size of hypotheses, via either syntactic restrictions or search strategies. This paper is concerned with polynomial induction and use of FOL hypotheses with no size restrictions. This is done via stochastic matching: instead of exhaustively exploring the set of matchings between any example and any short candidate hypothesis, one stochastically explores the set of matchings between any example and any candidate hypothesis. The user sets the number of matching samples to consider and thereby controls the cost of induction and classification. One advantage of this heuristic is to allow for resource-bounded learning, without any a priori knowledge about the problem domain. Experiments on a real-world problem pertaining to organic chemistry fully demonstrate the potentialities of the approach regarding both predictive accuracy and computational cost. 1
Tractable Induction and Classification in First Order Logic Via Stochastic Matching.
Learning in first-order logic (FOL) languages suffers from a specific difficulty: both induction and classification are potentially exponential in the size of hypotheses. This difficulty is usually dealt with by limiting the size of hypotheses, via either syntactic restrictions or search strategies. This paper is concerned with polynomial induction and use of FOL hypotheses with no size restrictions. This is done via stochastic matching: instead of exhaustively exploring the set of matchings between any example and any short candidate hypothesis, one stochastically explores the set of matchings between any example and any candidate hypothesis. The user sets the number of matching samples to consider and thereby controls the cost of induction and classification. One advantage of this heuristic is to allow for resource-bounded learning, without any a priori knowledge about the problem domain. Experiments on a real-world problem pertaining to organic chemistry fully demonstrate the potentialities of the approach regarding both predictive accuracy and computational cost. 1
Polynomial%2DTime Learning in Logic Programming and Constraint Logic Programming.
. Induction in first-order logic languages suffers from an additional factor of complexity compared to induction in attribute-value languages: the number of possible matchings between a candidate hypothesis and a training example. This paper investigates the use of a stochastic bias to control this factor of complexity: the exhaustive exploration of the matching space is replaced by considering a fixed number of matchings, obtained by random sampling. One thereby constructs from positive and negative examples a theory which is only approximately consistent. Both the degree of approximation and the computational cost of induction are controlled from the number of samples allowed. This approach is illustrated and validated on the mutagenesis problem. An ad hoc sampling mechanism has been purposely designed, and experimental results fully demonstrates the power of stochastic approximate induction, in terms of both predictive accuracy and computational cost. Furthermore, this approach appl...
Induction of Constraint Logic Programs.
. This paper provides a brief introduction and overview of the  emerging area of Inductive Constraint Logic Programming (ICLP). It  discusses some of the existing work in the area and presents some of the  research issues and open questions that need to be addressed.  1 Introduction  Inductive Logic Programming (ILP) refers to a class of machine learning algorithms where the agent learns a first-order theory from examples and background knowledge. The ILP framework in machine learning is perhaps the most general of all because of the complexity of the concepts learned. The use of first-order logic programs as the underlying representation makes ILP systems more powerful and useful than the conventional empirical machine learning systems. ILP systems have been successfully used in a variety of real life domains including mesh design, protein synthesis, games and fault diagnosis. Most of the first-order representations used in ILP systems are variants of horn-clause based clausal logic o...
A Knowledge%2DLevel Model of a Configurable Learning System.

Knowledge Level Model of a Configurable Learning System.

Flattening and Saturation Two Representation Changes for Generalization.

Approche logique de la g%E9n%E9ralisation en pr%E9sence de th%E9orie du domaine.

Panel Evaluating and Changing Representation in Concept Acquisition.

Semantic Model for Induction of First Order Theories.

Completeness for Inductive Procedures.
Abstract. Refinement equations play an important role in computer graphics and wavelet analysis. In this paper we investigate multivariate refinement equations associated with a dilation matrix and a finitely supported refinement mask. We characterize the Lp-convergence of a subdivision scheme in terms of the p-norm joint spectral radius of a collection of matrices associated with the refinement mask. In particular, the 2-norm joint spectral radius can be easily computed by calculating the eigenvalues of a certain linear operator on a finite dimensional linear space. Examples are provided to illustrate the general theory.
Saturation Postponing Choices when Inverting Resolution.

Beyond Inversion of Resolution.

