Digital Libraries %2D 6th Italian Research Conference

Approximate image color correlograms.

Coalition Structure Generation with GRASP.

A histogram%2Dbased technique for automatic threshold assessment in a run length smoothing%2Dbased algorithm.

Towards Learning to Rank in Description Logics.

Induction of Concepts in Web Ontologies through Terminological Decision Trees.

Relational Sequence based Classification in Multi%2Dagent Systems.

Using Explicit Word Co%2Doccurrences to Improve Term%2DBased Text Retrieval.

Merging Structural and Taxonomic Similarity for Text Retrieval Using Relational Descriptions.

A Multi%2Drelational Learning Approach for Knowledge Extraction in in Vitro Fertilization Domain.

Classifying Agent Behaviour through Relational Sequential Patterns.
Abstract. In Multi-Agent System, observing other agents and modelling their behaviour represents an essential task: agents must be able to quickly adapt to the environment and infer knowledge from other agents’ deportment. The observed data from this kind of environments are inherently sequential. We present a relational model to characterise adversary teams based on its behaviour using a set of relational sequences in order to classify them. We propose to use a relational learning algorithm to mine meaningful features as frequent patterns among the relational sequences and use these features to construct a feature vector for each sequence and then to compute a similarity value between sequences. The sequence extraction and classification are implemented in the domain of simulated robotic soccer, and experimental results are presented.
Learning to Rank Individuals in Description Logics Using Kernel Perceptrons.

Towards the induction of terminological decision trees.

Recovering uncertain mappings through structural validation and aggregation with the MoTo system.

Inductive reasoning and semantic web search.

Approximate Relational Reasoning by Stochastic Propositionalization.

Towards Multistrategic Statistical Relational Learning.
Abstract Statistical Relational Learning (SRL) is a growing field in Machine Learning that aims at the integration of logic-based learning approaches with probabilistic graphical models. Markov Logic Networks (MLNs) are one of the state-of-the-art SRL models that combine first-order logic and Markov networks (MNs) by attaching weights to first-order formulas and viewing these as templates for features of MNs. Learning models in SRL consists in learning the structure (logical clauses in MLNs) and the parameters (weights for each clause in MLNs). Structure learning of MLNs is performed by maximizing a likelihood function (or a function thereof) over relational databases and MLNs have been successfully applied to problems in relational and uncertain domains. However, most complex domains are characterized by incomplete data. Until now SRL models have mostly used Expectation-Maximization (EM) for learning statistical parameters under missing values. Multistrategic learning in the relational setting has been a successful approach to dealing with complex problems where multiple inference mechanisms can help solve different subproblems. Abduction is an inference strategy that has been proven useful for completing missing values in observations. In this paper we propose two frameworks for integrating abduction in SRL models. The first tightly integrates logical abduction with structure and parameter learning of MLNs in a single step. During structure search guided by conditional likelihood, clause evaluation is performed by first trying to logically abduce missing values in the data and then by learning optimal pseudo-likelihood parameters using the completed data. The second approach integrates abduction with Structural EM of [17] by performing logical abductive inference in the E-step and then by trying to maximize parameters in the M-step.
GRASP for the Coalition Structure Formation Problem
Abstract. The coalition structure formation problem represents an active research area in multi-agent systems. A coalition structure is defined as a partition of the agents involved in a system into disjoint coalitions. The problem of finding the optimal coalition structure is NP-complete. In order to find the optimal solution in a combinatorial optimization problem it is theoretically possible to enumerate the solutions and evaluate each. But this approach is infeasible since the number of solutions often grows exponentially with the size of the problem. In this paper we present a greedy adaptive search procedure (GRASP) to efficiently search the space of coalition structures in order to find an optimal one. Experiments and comparisons to other algorithms prove the validity of the proposed method in solving this hard combinatorial problem. 1
Feature Construction for Relational Sequence Learning
Abstract. Wetackletheproblemofmulti-classrelational sequencelearning using relevant patterns discovered from a set of labelled sequences. To deal with this problem, firstly each relational sequence is mapped into a feature vector using the result of a feature construction method. Since, the efficacy of sequence learning algorithms strongly depends on the features used to represent the sequences, the second step is to find an optimal subset of the constructed features leading to high classification accuracy. This feature selection task has been solved adopting a wrapper approach that uses a stochastic local search algorithm embedding a naïve Bayes classifier. The performance of the proposed method applied to a real-world dataset shows an improvement when compared to other established methods, such as hidden Markov models, Fisher kernels and conditional random fields for relational sequences.
Fuzzy Clustering for Semantic Knowledge Bases.

Inductive learning for the Semantic Web What does it buy%3F

Plugging Taxonomic Similarity in First%2DOrder Logic Horn Clauses Comparison.

High Performing Algorithms for MAP and Conditional Inference in Markov Logic.

Relational Temporal Data Mining for Wireless Sensor Networks.

Intelligent Text Processing Techniques for Textual%2DProfile Gene Characterization.

ReduCE A Reduced Coulomb Energy Network Method for Approximate Classification.

A Distance%2DBased Technique for Non%2DManhattan Layout Analysis.
Layout analysis is a fundamental step in automatic document processing. Many different techniques have been proposed in literature to perform this task. These are broadly divided in two main categories according to the approach they follow: the top-down methods start by identifying the high level components of the page structure and then recursively split them until basic blocks are found. On the other hand, bottom-up approaches start with the smallest elements (e.g., the pixels in case of digitized document) and then recursively merge them into higher level components. A first limitation of such methods is that most of them are designed to deal only with digitized documents and hence are not applicable to the general case of native digital documents which are mainly diffused. Furthermore, top-down and most of bottom-up methods are able to process Manhattan layout documents only. In this work, we propose a general bottom-up strategy to tackle the layout analysis of (possibly) non-Manhattan documents, and two specializations of it to handle the different cases of bitmaps and PS/PDF sources. 1
A Logic Programming Framework for Learning by Imitation.
Humans use imitation as a mechanism for acquiring knowledge, i.e. they use instructions and/or demonstrations provided by other humans. In this paper we propose a logic programming framework for learning from imitation in order to make an agent able to learn from relational demonstrations. In particular, demonstrations are received in incremental way and used as training examples while the agent interacts in a stochastic environment. This logical framework allows to represent domain specific knowledge as well as to compactly and declaratively represent complex relational processes. The framework has been implemented and validated with experiments in simulated agent domains. 1
A Logic Programming Framework for Learning by Imitation.
Humans use imitation as a mechanism for acquiring knowledge, i.e. they use instructions and/or demonstrations provided by other humans. In this paper we propose a logic programming framework for learning from imitation in order to make an agent able to learn from relational demonstrations. In particular, demonstrations are received in incremental way and used as training examples while the agent interacts in a stochastic environment. This logical framework allows to represent domain specific knowledge as well as to compactly and declaratively represent complex relational processes. The framework has been implemented and validated with experiments in simulated agent domains. 1
Nonmonotonic Onto%2DRelational Learning.

Inductive Query Answering and Concept Retrieval Exploiting Local Models.

Fuzzy Clustering for Categorical Spaces.
Abstract. A multi-relational clustering method is presented which can be applied to complex knowledge bases storing resources expressed in the standard Semantic Web languages. It adopts effective and language-independent dissimilarity measures that are based on a finite number of dimensions corresponding to a committee of discriminating features(represented by concept descriptions). The clustering algorithm expresses the possible clusterings in tuples of central elements (medoids, w.r.t. the given metric) of variable length. It iteratively adjusts these centers following the rationale of fuzzy clustering approach, i.e. one where the membership to each cluster is not deterministic but rather ranges in the unit interval. An experimentation with some ontologies proves the feasibility of our method and its effectiveness in terms of clustering validity indices. 1 Clustering in Complex Domains Recently, multi-relational learning methods are being devised for knowledge bases in the Semantic Web (henceforth SW), expressed in the standard representations. Indeed, the most burdensome related maintenance tasks, such as ontology construction, refinement
Relational Sequence Clustering for Aggregating Similar Agents.
Abstract. Many clustering methods are based on flat descriptions, while data regarding real-world domains include heterogeneous objects related to each other in multiple ways. For instance, in the field of Multi-Agent System, multiple agents interact with the environment and with other agents. In this case, in order to act effectively an agent should be able to recognise the behaviours adopted by other agents. Actions taken by an agent are sequential, and thus its behaviour can be expressed as a sequence of actions. Inferring knowledge about competing and/or companion agents by observing their actions is very beneficial to construct a behavioural model of the agent population. In this paper we propose a clustering method for relational sequences able to aggregate companion agent behaviours. The algorithm has been tested on a real world dataset proving its validity.
Efficient MAP Inference for Statistical Relational Models through Hybrid Metaheuristics.

Relational Learning by Imitation.
Abstract. Imitative learning can be considered an essential task of humans development. People use instructions and demonstrations provided by other human experts to acquire knowledge. In order to make an agent capable of learning through demonstrations, we propose a relational framework for learning by imitation. Demonstrations and domain specific knowledge are compactly represented by a logical language able to express complex relational processes. The agent interacts in a stochastic environment and incrementally receives demonstrations. It actively interacts with the human by deciding the next action to execute and requesting demonstration from the expert based on the current learned policy. The framework has been implemented and validated with experiments in simulated agent domains.
Evidential Nearest%2DNeighbors Classification for Inductive ABox Reasoning.
In the line of our investigation on inductive methods for Semantic Web reasoning, we propose an alternative way for approximate ABox reasoning based on the analogical principle of the nearestneighbors. Once neighbors of a test individual are selected, a combination rule descending from the Dempster-Shafer theory can join together the evidence provided by the neighbor individuals. We show how to exploit the procedure for determining unknown class- and role-memberships or fillers for datatype properties which may be the basis for many further ABox inductive reasoning algorithms.
Approximate Classification of Semantically Annotated Web Resources Exploiting Pseudo%2Dmetrics Induced by Local Models.

A Semantic Similarity Measure for Expressive Description Logics
Abstract. A totally semantic measure is presented which is able to calculate a similarity value between concept descriptions and also between concept description and individual or between individuals expressed in an expressive description logic. It is applicable on symbolic descriptions although it uses a numeric approach for the calculus. Considering that Description Logics stand as the theoretic framework for the ontological knowledge representation and reasoning, the proposed measure can be effectively used for agglomerative and divisional clustering task applied to the semantic web domain. 1
A General Similarity Framework for Horn Clause Logic.

Metric%2Dbased stochastic conceptual clustering for ontologies.

Inductive Classification of Semantically Annotated Resources through Reduced Coulomb Energy Networks.

Post%2Dproceedings of the Forth Italian Research Conference on Digital Library Systems

Evolutionary Clustering in Description Logics Controlling Concept Formation and Drift in Ontologies.

Evolutionary Clustering in Description Logics Controlling Concept Formation and Drift in Ontologies.

Structure Learning of Markov Logic Networks through Iterated Local Search.

Query Answering and Ontology Population An Inductive Approach.
Abstract. In order to overcome the limitations of deductive logic-based approaches to deriving operational knowledge from ontologies, especially when data come from distributed sources, inductive (instance-based) methods may be better suited, since they are usually efficient and noisetolerant. In this paper we propose an inductive method for improving the instance retrieval and enriching the ontology population. By casting retrieval as a classification problem with the goal of assessing the individual class-memberships w.r.t. the query concepts, we propose an extension of the k-Nearest Neighbor algorithm for OWL ontologies based on an entropic distance measure. The procedure can classify the individuals w.r.t. the known concepts but it can also be used to retrieve individuals belonging to query concepts. Experimentally we show that the behavior of the classifier is comparable with the one of a standard reasoner. Moreover we show that new knowledge (not logically derivable) is induced. It can be suggested to the knowledge engineer for validation, during the ontology population task. 1
Conceptual Clustering and Its Application to Concept Drift and Novelty Detection.

k%2DNearest Neighbor Classification on First%2DOrder Logic Descriptions.

Incremental machine learning techniques for document layout understanding.

Non%2Dparametric Statistical Learning Methods for Inductive Classifiers in Semantic Knowledge Bases.
This work concerns non-parametric approaches for statistical learning applied to the standard knowledge representations languages adopted in the Semantic Web context. We present methods based on epistemic inference that are able to elicit the semantic similarity of individuals in OWL knowledge bases. Specifically, a totally semantic and language independent semi-distance function is presented and from it, an epistemic kernel function for Semantic Web representations is derived. Both the measure and the kernel function are embedded into non-parametric statistical learning algorithms customized for coping with Semantic Web representations. Particularly, the measure is embedded into a k-Nearest Neighbor algorithm and the kernel function is embedded in a Support Vector Machine. The realized algorithms are used to performe inductive concept retrieval and query answering. An experimentation on real ontologies proves that the methods can be effectively employed for performing the target tasks and moreover that it is possible to induce new assertions that are not logically derivable. 1 Learning from Ontologies The Semantic Web (SW) represents an emerging domain where business, enterprise and organization on the Web will have its own organizational model (an ontology), and knowledge intensive automated manipulations on complex relational descriptions are foreseen. Specific formal languages for knowledge representation have been designed for supporting a variety of applications in this context spanning from biological and geospatial fields to agents technology and service oriented architectures. Description Logics (DLs) [1], a family of languages that is endowed with well-founded semantics and reasoning services, have been adopted as the core of the ontology language (OWL 1). 1
DL%2DFOIL Concept Learning in Description Logics.

Foundations of Onto%2DRelational Learning.

Learning with Kernels in Description Logics.
Abstract. We tackle the problem of statistical learning in the standard knowledge base representations for the Semantic Web which are ultimately expressed in description Logics. Specifically, in our method a kernel functions for the ALCN logic integrates with a support vector machine which enables the usage of statistical learning with reference representations. Experiments where performed in which kernel classification is applied to the tasks of resource retrieval and query answering on OWL ontologies. 1 Learning from Ontologies The Semantic Web (SW) represents an emerging applicative domain where knowledge intensive automated manipulations on complex relational descriptions are foreseen. Although machine learning techniques may have a great potential in this field, so far research has focused mainly on methods for knowledge acquisition from text (ontology learning) [4]. Yet machine learning methods can be transposed from ILP to be applied to ontologies described with formal concept
Discriminative Structure Learning of Markov Logic Networks.

Automatic Document Organization Exploiting FOL Similarity%2Dbased Techniques.

A Multi%2Drelational Hierarchical Clustering Method for DatalogKnowledge Bases.

Stochastic Propositionalization for Efficient Multi%2Drelational Learning.
Abstract. The efficiency of multi-relational data mining algorithms, addressing the problem of learning First Order Logic (FOL) theories, strongly depends on the search method used for exploring the hypotheses space and on the coverage test assessing the validity of the learned theory against the training examples. A way of tackling the complexity of this kind of learning systems is to use a propositional method that reformulates a multi-relational learning problem into an attribute-value one. We propose a population based algorithm that using a stochastic propositional method efficiently learns complete FOL definitions. 1
Statistical Learning for Inductive Query Answering on OWL Ontologies.

Learning Semantic Web Rules within the Framework of SHIQ log.
In this paper we face the problem of learning Semantic Web rules within a decidable instantiation of the DL+log framework which integrates the DL SHIQ and positive Datalog. To solve the problem, we resort to the methodological apparatus of Inductive Logic Programming.
Distance%2DBased Classification in OWL Ontologies.

Approximate Reasoning for Efficient Anytime Induction from Relational Knowledge Bases.
relational knowledge bases
A Note on the Evaluation of Inductive Concept Classification Procedures.
Abstract. The limitations of deductive logic-based approaches at deriving operational knowledge from ontologies may be overcome by inductive (instancebased) methods, which are usually efficient and noise-tolerant. However the evaluation of such methods is made particularly difficult by the open-world semantics which may often cause individuals not to be deductively classified by the reasoner. In this paper an evaluation method is proposed that is suitable for comparing inductive classification methods to standard reasoners. Experimentally we show that the behavior of a nearest neighbor classifier is comparable with the one of a standard reasoner in terms of the proposed indices. 1
Learning SHIQ log Rules for Ontology Evolution.
Abstract. The definition of new concepts or roles for which extensional knowledge become available can turn out to be necessary to make a DL ontology evolve. In this paper we reformulate this task as a machine learning problem and study a solution based on techniques borrowed from that form of logic-based machine learning known under the name of Inductive Logic Programming (ILP). More precisely, we propose to adapt previous ILP results to the knowledge representation framework of DL+log in order to learn rules to be used for changing SHIQ ontologies. 1
Representing Uncertain Concepts in Rough Description Logics via Contextual Indiscernibility Relations.
Abstract. We investigate on modeling uncertain concepts via rough description logics, which are an extension of traditional description logics by a simple mechanism to handle approximate concept definitions through lower and upper approximations of concepts based on a rough-set semantics. This allows to apply rough description logics for modeling uncertain knowledge. Since these approximations are ultimately grounded on an indiscernibility relationship, the paper explores possible logical and numerical ways for defining such relationships based on the considered knowledge. In particular, the notion of context is introduced, allowing for the definition of specific equivalence relationships, to be used for approximations as well as for determining similarity measures, which may be exploited for introducing a notion of tolerance in the indiscernibility. 1
Analogical Reasoning in Description Logics.
This work presents a framework, founded on multi-relational instancebased learning, for inductive (memory-based) reasoning on knowledge bases expressed in Description Logics. The procedure, which exploits a relational dissimilarity measure based on the notion of Information Content, can be employed both to answer to class-membership queries and to predict assertions, that may not be logically entailed by the knowledge base. These tasks may be the baseline for other inductive methods for ontology construction and evolution. In a preliminary experimentation, we show that the method is sound. Besides it is actually able to induce new knowledge that might be acquired in the knowledge base.
Approximate Measures of Semantic Dissimilarity under Uncertainty.
Abstract. We propose semantic distance measures based on the criterion of approximate discernibility and on evidence combination. In the presence of incomplete knowledge, the distance measures the degree of belief in the discernibility of two individuals by combining estimates of basic probability masses related to a set of discriminating features. We also suggest ways to extend this distance for comparing individuals to concepts and concepts to other concepts. 1
Machine Learning for Digital Document Processing from Layout Analysis to Metadata Extraction.
Summary. In the last years, the spread of computers and the Internet caused a significant amount of documents to be available in digital format. Collecting them in digital repositories raised problems that go beyond simple acquisition issues, and cause the need to organize and classify them in order to improve the effectiveness and efficiency of the retrieval procedure. The success of such a process is tightly related to the ability of understanding the semantics of the document components and content. Since the obvious solution of manually creating and maintaining an updated index is clearly infeasible, due to the huge amount of data under consideration, there is a strong interest in methods that can provide solutions for automatically acquiring such a knowledge. This work presents a framework that intensively exploits intelligent techniques to support different tasks of automatic document processing from acquisition to indexing, from categorization to storing and retrieval. The prototypical version of the system DOMINUS is presented, whose main characteristic is the use of a Machine Learning Server, a suite of different inductive learning methods and systems, among which the more suitable for each specific document processing phase is chosen and applied. The core system is the incremental first-order logic learner INTHELEX. Thanks to incrementality, it can continuously update and refine the learned theories, dynamically extending its knowledge to handle even completely new classes of documents. Since DOMINUS is general and flexible, it can be embedded as a document management engine into many different Digital Library systems. Experiments in a real-world domain scenario, scientific conference management, confirmed the good performance of the proposed prototype. 1
Multi%2DDimensional Relational Sequence Mining.
 The issue addressed in this paper concerns the discovery of frequent multi-dimensional patterns from relational sequences. The great variety of applications of sequential pattern mining, such as user profiling, medicine, local weather forecast and bioinformatics, makes this problem one of the central topics in data mining. Nevertheless, sequential information may concern data on multiple dimensions and, hence, the mining of sequential patterns from multi-dimensional information results very important. In a multi-dimensional sequence each event depends on more than one dimension, such as in spatio-temporal sequences where an event may be spatially or temporally related to other events. In literature, the multi-relational data mining approach has been successfully applied to knowledge discovery from complex data. However, there exists no contribution to manage the general case of multi-dimensional data in which, for example, spatial and temporal information may co-exist. This work takes into account the possibility to mine complex patterns, expressed in a first-order language, in which events may occur along different dimensions. Specifically, multidimensional patterns are defined as a set of atomic first-order formulae in which events are explicitly represented by a variable and the relations between events are represented by a set of dimensional predicates. A complete framework and an Inductive Logic Programming algorithm to tackle this problem are presented along with some experiments on artificial and real multi-dimensional sequences proving its effectiveness.   
Induction of Classifiers through Non%2DParametric Methods for Approximate Classification and Retrieval with Ontologies.

Evolutionary Conceptual Clustering Based on Induced Pseudo%2DMetrics.

Post%2Dproceedings of the Third Italian Research Conference on Digital Library Systems

Instance%2DBased Query Answering with Semantic Knowledge Bases.
Abstract. A procedure founded in instance-based learning is presented, for performing a form of analogical reasoning on knowledge bases expressed in a wide range of ontology languages. The procedure exploits a novel semi-distance measure for individuals, that is based on their semantics w.r.t. a number of dimensions corresponding to a committee of features represented by concept descriptions. The procedure can answer by analogy to class’membership queries on the grounds of the classification of a number of training instances (the nearest ones w.r.t. the semi-distance measure). Particularly, it may also predict assertions that are not logically entailed by the knowledge base. In the experimentation, where we compare the procedure to a logical reasoner, we show that it can be quite accurate and augment the scope of its applicability, outperforming previous prototypes that adopted other semantic measures. 1
A Hierarchical Clustering Procedure for Semantically Annotated Resources.
Abstract. A clustering method is presented which can be applied to relational knowledge bases. It can be used to discover interesting groupings of resources through their (semantic) annotations expressed in the standard languages employed for modeling concepts in the Semantic Web. The method exploits a simple (yet effective and language-independent) semi-distance measure for individuals, that is based on the resource semantics w.r.t. a number of dimensions corresponding to a committee of features represented by a group of concept descriptions (discriminating features). The algorithm is an fusion of the classic Bisecting k-Means with approaches based on medoids since they are intended to be applied to relational representations. We discuss its complexity and the potential applications to a variety of important tasks. 1 Learning Methods for Concept Languages In the inherently distributed applications related to the Semantic Web (henceforth SW) there is an extreme need of automatizing those activities which are
Similarity%2DGuided Clause Generalization.

Randomized metric induction and evolutionary conceptual clustering for semantic knowledge bases.
We present an evolutionary clustering method which can be applied to multi-relational knowledge bases storing resource annotations expressed in the standard languages for the Semantic Web. The method exploits an effective and languageindependent semi-distance measure defined for the space of individual resources, that is based on a finite number of dimensions corresponding to a committee of discriminating features (represented by concept descriptions). A maximally discriminating group of features can be obtained with the randomized optimization methods described in the paper. The clustering algorithm represents the possible clusterings as strings of central elements (medoids, w.r.t. the given metric) of variable length. Hence, the number of clusters is not required as a parameter since the method is able to find an optimal choice by means of the evolutionary operators and of a proper fitness function. We also show how to assign each cluster with a newly constructed intensional definition in the employed concept language. An experimentation with some ontologies proves the feasibility of our method and its effectiveness in terms of clustering validity indices.
Induction of Optimal Semi%2Ddistances for Individuals based on Feature Sets.
Abstract. Many activities related to semantically annotated resources can be enabled by a notion of similarity among them. We propose a method for defining a family of semi-distances over the set of individuals in a knowledge base which can be used in these activities. In the line of works on distance-induction on clausal spaces, the family is parameterized on a committee of concepts. Hence, we also present a method based on the idea of simulated annealing to be used to optimize the choice of the best concept committee. 1
Incremental Learning of First Order Logic Theories for the Automatic Annotations of Web Documents.
Organizing large repositories spread throughout the most diverse Web sites rises the problem of effective storage and efficient retrieval of documents. This can be obtained by selectively extracting from them the significant textual information, contained in peculiar layout components, that in turn depend on the identification of the correct document class. The continuous flow of new and different documents in a weakly structured environment like the Web calls for incrementality, as the ability to continuously update or revise a faulty knowledge previously acquired, while the need to express structural relations among layout components suggest the exploitation of a powerful and symbolic representation language. This paper proposes the application of incremental first-order logic learning techniques in the document layout preprocessing steps, supported by good results obtained in experiments on a real dataset. 1
Evolutionary Conceptual Clustering of Semantically Annotated Resources.
A clustering method is presented which can be applied to knowledge bases storing semantically annotated resources. The method can be used to discover groupings of structured objects expressed in the standard concept languages employed in the Semantic Web. The method exploits effective language-independent semi-distance measures over the space of resources. These are based on their semantics w.r.t. a number of dimensions corresponding to a committee of features represented by a group of discriminating concept descriptions. We show how to obtain a maximally discriminating group of features through a feature construction procedure based on genetic programming. The evolutionary clustering algorithm employed is based on the notion of medoids applied to relational representations. It is able to induce an optimal set of clusters by means of a proper fitness function based on the defined distance and the discernibility criterion. An experimentation with some real ontologies proves the feasibility of our method. 1.
Unsupervised Discretization Using Kernel Density Estimation.

Induction of Optimal Semantic Semi%2Ddistances for Clausal Knowledge Bases.
Abstract. Several activities related to semantically annotated resources can be enabled by a notion of similarity, spanning from clustering to retrieval, matchmaking and other forms of inductive reasoning. We propose the definition of a family of semi-distances over the set of objects in a knowledge base which can be used in these activities. In the line of works on distance-induction on clausal spaces, the family is parameterized on a committee of concepts expressed with clauses. Hence, we also present a method based on the idea of simulated annealing to be used to optimize the choice of the best concept committee. 1
Generalization%2DBased Similarity for Conceptual Clustering.

Conceptual Clustering Applied to Ontologies.

Instance%2Dbased retrieval by analogy.
This work presents a method for retrieval in knowledge bases expressed in Description Logics, founded in the instancebased learning. The procedure implements the disjunctive version space approach exploiting a notion of semantic difference. The method can be employed both to answer to class-membership queries, even though the answers are not logically entailed by the knowledge base, e.g. there are some inconsistent assertions due to heterogeneous sources. In addition, it may also predict/suggest new assertions The method has been implemented and tested in an experimentation, where we show that it is sound and effective.
Approximate Query Answering and Ranking for Semantic Knowledge Bases.

Mining Time%2Dseries Sequences of Reactions for Biologically Active Patterns in Metabolic Pathways.

A Fast Partial Memory Approach to Incremental Learning through an Advanced Data Storage Framework.

Efficient Discovery of Services Specified in Description Logics Languages.
Semantic service descriptions are frequently given using expressive ontology languages based on description languages. The expressiveness of these languages, however, often implies problems for efficient service discovery, especially when increasing numbers of services become available in large organizations and on the Web. To remedy this problem, we propose an efficient service discovery/retrieval method grounded on a conceptual clustering approach, where services are specified in Description Logics as class definitions [10] and they are retrieved by defining a class expression as a query and by computing the individual subsumption relationship between the query and the available descriptions. We present a new conceptual clustering method that constructs tree indices for clustered services, where available descriptions are the leaf nodes, while inner nodes are intensional descriptions (generalization) of their children nodes. The matchmaking is performed by following the tree branches whose nodes might satisfy the query. The query answering time may strongly improve, since the number of retrieval steps may decrease from O(n) to O(log n) for concise queries. We also show that the proposed method is sound and complete.
Building Rules on top of Ontologies%3F Inductive Logic Programming can help%21

Semantic Nearest Neighbor Search in OWL Ontologies.

Approximate Measures of Semantic Dissimilarity under Uncertainty.
Abstract. We propose semantic distance measures based on the criterion of approximate discernibility and on evidence combination. In the presence of incomplete knowledge, the distance measures the degree of belief in the discernibility of two individuals by combining estimates of basic probability masses related to a set of discriminating features. We also suggest ways to extend this distance for comparing individuals to concepts and concepts to other concepts. 1
Multi%2Dclass Protein Fold Recognition Through a Symbolic%2DStatistical Framework.

Introduction to the special issue.

Inference of abduction theories for handling incompleteness in first%2Dorder learning.
Inference of abduction theories for handling
Foundations of Intelligent Systems
The aim of this paper is to introduce the theoretical foundations of an approach for intelligent systems development. Derived from semiotics, a classic discipline in human sciences, the theory developed provides a mathematical framework for the concept of knowledge and for knowledge processing. As a result, a new perspective to study and to develop intelligent systems emerges. A taxonomy of elementary types of knowledge is proposed based on the classification of types of signs in semiotics, followed by a another classification of knowledge from the point of view of application in cognitive systems. In addition, we propose the mathematical definition of objects, objects systems and objects networks, to model mathematically the different types of knowledge described. The symbiosis of such key concepts introduces a computational paradigm to develop and implement intelligent systems, called here computational semiotics. 
Automatic Content%2Dbased Indexing of Digital Documents through Intelligent Processing Techniques.
The availability of large, heterogeneous repositories of electronic documents is increasing rapidly, and the need for flexible, sophisticated document manipulation tools is growing correspondingly. This work presents DOMINUS a system for automated electronic documents processing characterized by the intensive exploitation of intelligent techniques in each step of the process from document acquisition in electronic format to document indexing for categorization and information retrieval purposes. It embeds incremental learning techniques useful in the context of web libraries where documents are acquired in an incremental fashion. Since the system is general and flexible, it can be embedded as a document management engine into many different domain-specific applications. Here, we present the exploitation of the system on the Conference Management domain, and show how DOMINUS can usefully support some of the more critical and knowledge-intensive tasks involved by the organization of a scientific conference.  
Two Orthogonal Biases for Choosing the Intensions of Emerging Concepts in Ontology Refinement.
FAIA covers all aspects of theoretical and applied artificial intelligence research in the form of monographs, doctoral dissertations, textbooks, handbooks and proceedings volumes. The FAIA series contains several sub-series, including “Information Modelling and Knowledge Bases ” and
Tractable Feature Generation Through Description Logics with Value and Number Restrictions.
Abstract. In the line of a feature generation paradigm based on relational concept descriptions, we extend the applicability to other languages of the Description Logics family endowed with specific language constructors that do not have a counterpart in the standard relational representations, such as clausal logics. We show that the adoption of an enhanced language does not increase the complexity of feature generation, since the process is still tractable. Moreover this can be considered as a formalization for future employment of even more expressive languages from the Description Logics family. 1
Automatic Topics Identification for Reviewer Assignment.

On the Missing Link Between Frequent Pattern Discovery and Concept Formation.

Towards Automatic Digital Library Content%2Dbased Management.

A dissimilarity measure for ALC concept descriptions.

Classifying Aggregated Data a Symbolic Data Analysis Approach.


Reasoning by Analogy in Description Logics Through Instance%2Dbased Learning.
Abstract — This work presents a method founded in instancebased learning for inductive (memory-based) reasoning on ABoxes. The method, which exploits a semantic dissimilarity measure between concepts and instances, can be employed both to answer class membership queries and to predict new assertions that may be not logically entailed by the knowledge base. In a preliminary experimentation, we show that the method is sound and it is actually able to induce new assertions that might be acquired in the knowledge base. I.
Analogical Reasoning in Description Logics.
This work presents a framework, founded on multi-relational instancebased learning, for inductive (memory-based) reasoning on knowledge bases expressed in Description Logics. The procedure, which exploits a relational dissimilarity measure based on the notion of Information Content, can be employed both to answer to class-membership queries and to predict assertions, that may not be logically entailed by the knowledge base. These tasks may be the baseline for other inductive methods for ontology construction and evolution. In a preliminary experimentation, we show that the method is sound. Besides it is actually able to induce new knowledge that might be acquired in the knowledge base.

Classification of symbolic objects A lazy learning approach.

Machine Learning and Data Mining.
this article, we focus primarily on the role of machine learning algorithms in the data mining process. The above medical example thus represents a prototypical data mining problem in which the data consists of a collection of time series descriptions, and we use the data to learn to predict later events in the series (emergency C-section) based on earlier events (symptoms before delivery). Although we used a medical example for illustration, we could have given an analogous example of learning to predict which bank loan applicants are at high risk of failing to repay their loan (see Figure 2). As shown in the figure, in this application data typically consists of time series descriptions of customers&#039; bank balances and other demographic information, rather than medical symptoms. Yet other applications, shown in Figure 3, include predicting customer purchase behavior, customer retention, and the quality of goods produced on a particular manufacturing line. All of these are applications for which data mining has been successfully applied in practice, and where further research promises yet more effective techniques. 3 The State of the Art, and What Next?
Innovations in Applied Artificial Intelligence
Mutual adaptation is an important phenomenon in human-human communications. Traditionally HRI research was more interested in investigating adaptation of the robot to the human using machine learning techniques but the possibility of utilizing the natural ability of humans to adapt to other humans and artifacts including robots is recently becoming increasingly attractive. This paper presents some of the results from an experiment conducted to investigate the interaction patterns and effectiveness of motion cues as a feedback modality between a human operator and a miniature robot in a confined collaborative navigation task. The results presented in this paper show evidence of human adaptation to the robot and moreover suggest that the adaptation rate is not constant or continuous in time but is discontinuous and nonlinear. The results also show evidence of a starting exploration stage before the adaptation with duration dependent on the expectations of the human regarding the capabilities of the robot in the given task. The paper investigates how to utilize these and related findings for building robots not only capable of adapting to human operators but can also help those operators adapt to them. I.
Avoiding Order Effects in Incremental Learning.
This paper addresses the problem of mitigating the order  e#ects in incremental learning, a phenomenon observed when di#erent  ordered sequences of observations lead to di#erent results. A modification  of an ILP incremental learning system, with the aim of making it  order-independent, is presented. A backtracking strategy on theories is  incorporated in its refinement operators, which causes a change of its  refinement strategy and reflects the human behavior during the learning  process. A modality to restore a previous theory, in order to backtrack  on a previous knowledge level, is presented. Experiments validate the  approach in terms of computational cost and predictive accuracy.
Intelligent Document Processing.
Digital repositories raise the need for an effective and efficient retrieval of the stored material. In this paper we propose the intensive application of intelligent techniques to the steps of document layout analysis, document image classification and understanding on digital documents. Specifically, the complex interrelation existing among layout components, that are fundamental to assign them the proper semantic role, suggest the exploitation of first-order representations in some learning steps. Results obtained in a prototypical system for scientific conference management prove that the proposed approach can be beneficial both for the layout recognition and for the selection of interesting components of the document, from which extracting the text for categorizing the document according to its topic. 1.
A Counterfactual%2DBased Learning Algorithm for Description Logic.

Handling Continuous%2DValued Attributes in Incremental First%2DOrder Rules Learning.
Abstract. Machine Learning systems are often distinguished according to the kind of representation they use, which can be either propositional or first-order logic. The framework working with first-order logic as a representation language for both the learned theories and the observations is known as Inductive Logic Programming (ILP). It has been widely shown in the literature that ILP systems have limitations in dealing with large amounts of numerical information, that is however a peculiarity of most real-world application domains. In this work we present a strategy to handle such information in a relational learning incremental setting and its integration with classical symbolic approaches to theory revision. Experiments were carried out on a real-world domain and a comparison with a state-of-art system is reported. 1
On the LearnAbility of Abstraction Theories from Observations for Relational Learning.
Abstract. The most common methodology in symbolic learning consists in inducing, given a set of observations, a general concept definition. It is widely known that the choice of the right description language for a learning problem can affect the efficacy and effectiveness of the learning task. Furthermore, most of the real-world domain are contaminated by various kinds of imperfections in data such as inappropriateness of the description language which does not contain/facilitate an exact representation of the target concept. To deal with such kind of situations, Machine Learning approaches have moving from a framework exploiting a single inference mechanism, such as induction, towards one integrating multiple inference strategies such as abstraction. The literature so far assumed that the information needed to the learning systems to apply additional inference strategies is provided by the an expert domain. The objective in this work is the automatic inference of such information. The efficacy of the proposed method in generating effective theories to perform abstraction was tested by providing the generated abstraction theories to the learning system INTHELEX allowing it to exploit its multistrategy capabilities, in particular the abstraction one. Various experiments were carried out on a real-world application domain of scientific paper documents showing the validity of the proposed method. 1
REDD An Algorithm for Redundancy Detection in RDF Models.

Intelligent Document Processing.
Digital repositories raise the need for an effective and efficient retrieval of the stored material. In this paper we propose the intensive application of intelligent techniques to the steps of document layout analysis, document image classification and understanding on digital documents. Specifically, the complex interrelation existing among layout components, that are fundamental to assign them the proper semantic role, suggest the exploitation of first-order representations in some learning steps. Results obtained in a prototypical system for scientific conference management prove that the proposed approach can be beneficial both for the layout recognition and for the selection of interesting components of the document, from which extracting the text for categorizing the document according to its topic. 1.
Automatic Induction of Abduction and Abstraction Theories from Observations.
Traditional Machine Learning approaches are based on single  inference mechanisms. A step forward concerned the integration of  multiple inference strategies within a first-order logic learning framework,  taking advantage of the benefits that each approach can bring.
ILP Meets Knowledge Engineering A Case Study.

Mining the Semantic Web A Logic%2DBased Methodology.

Semantic%2DBased Access to Digital Document Databases.
Discovering significant meta-information from document collections  is a critical factor for knowledge distribution and preservation.
A Dissimilarity Measure for the ALC Description Logic.
Abstract. This work presents a dissimilarity measure for an expressive Description Logic endowed with the principal constructors employed in the standard representations for ontological knowledge. In particular, the focus is on the definition of a dissimilarity measure for the ALC description logic based both on the syntax and on the semantics of the descriptions. The measure is shown to be applicable to assess the dissimilarity for cases involving individuals. 1
An ILP Perspective on the Semantic Web.

Machine Learning ECML 2004

Knowledge Discovery in Databases PKDD 2004

Machine Learning Methods for Automatically Processing Historical Documents From Paper Acquisition to XML Transformation.
One of the aims of the EU project COLLATE is to design and implement a Web-based collaboratory for archives, scientists and end-users working with digitized cultural material. Since the originals of such a material are often unique and scattered in various archives, severe problems arise for their wide fruition. A solution would be to develop intelligent document processing tools that automatically transform printed documents into a webaccessible form such as XML. Here, we propose the use of a document processing system, WISDOM++, which uses heavily machine learning techniques in order to perform such a task, and report promising results obtained in preliminary experiments.
Induction and Revision of Terminologies.

Automatic Induction of Domain%2DRelated Information Learning Descriptors Type Domains.

A Backtracking Strategy for Order%2DIndependent Incremental Learning.
Agents that exist in an environment that changes over time, and are able to take into account the temporal nature of experience, are commonly called incremental learners. It is widely known that incremental learning systems suffer from order effects, a phenomenon observed when differently ordered sequences of examples lead to different results. The goal of this paper is presenting INTHELEXback , an order-independent evolution of the incremental learning system INTHELEX. A backtracking strategy is incorporated in its refinement operators, which causes a change in its refinement strategy and reflects the human behavior during the learning process. It consists in remembering the different versions of the learned theory across modifications due to new evidence. In this way the system can backtrack on a previous knowledge level when it discovers to have made a wrong choice. Experiments on an artificial dataset validate the approach in terms of computational cost and predictive accuracy.
Personalizing Digital Libraries for Education.

An Algorithm for Incremental Mode Induction.

Incremental Induction of Classification Rules for Cultural Heritage Documents.

Machine Learning Approaches for Inducing Student Models.
The main issue in e-learning is student modelling, i.e. the analysis of  a student&#039;s behaviour and prediction of his/her future behaviour and learning  performance. Indeed, it is difficult to monitor the students&#039; learning behaviours.
Automatic Induction of First%2DOrder Logic Descriptors Type Domains from Observations.

Top%2DDown Induction of Model Trees with Regression and Splitting Nodes.
Model trees are an extension of regression trees that associate leaves with multiple regression models. In this paper, a  method for the data-driven construction of model trees is presented, namely, the Stepwise Model Tree Induction (SMOTI) method. Its  main characteristic is the induction of trees with two types of nodes: regression nodes, which perform only straight-line regression, and  splitting nodes, which partition the feature space. The multiple linear model associated with each leaf is then built stepwise by  combining straight-line regressions reported along the path from the root to the leaf. In this way, internal regression nodes contribute to  the definition of multiple models and have a &#034;global&#034; effect, while straight-line regressions at leaves have only &#034;local&#034; effects.
Efficient Evaluation of Candidate Hypotheses in AL%2Dlog.

Discovering Logical Structures in Digital Documents.

Knowledge%2DIntensive Induction of Terminologies from Metadata.

Top%2DDown Induction of Model Trees with Regression and Splitting Nodes.
Model trees are an extension of regression trees that associate leaves with multiple regression models. In this paper, a  method for the data-driven construction of model trees is presented, namely, the Stepwise Model Tree Induction (SMOTI) method. Its  main characteristic is the induction of trees with two types of nodes: regression nodes, which perform only straight-line regression, and  splitting nodes, which partition the feature space. The multiple linear model associated with each leaf is then built stepwise by  combining straight-line regressions reported along the path from the root to the leaf. In this way, internal regression nodes contribute to  the definition of multiple models and have a &#034;global&#034; effect, while straight-line regressions at leaves have only &#034;local&#034; effects.
Incremental learning and concept drift in INTHELEX.

Discovering Student Models in e%2DLearning Systems.
In all areas of the e-era, personalization plays an important role. Particularly in elearning a main issue is student modeling, that is the analysis of student behavior and prediction of his/her future behavior and learning performance. In fact, nowadays, the most prevailing issue in the e-learning environment is that it is not easy to monitor students&#039; learning behaviors. In this paper we have focused our attention on the system (the Profile Extractor) based on Machine Learning techniques, which allows for the discovery of preferences, needs and interests of users that have access to an e-learning system. The automatic generation and the discovery of the user profile, to agree as simple student model based on the learning performance and the communication preferences, allow creating a personalized education environment. Moreover, we presented an evaluation of the accuracy of the Profile Extractor system using the classical Information Retrieval metrics.
A Complete Subsumption Algorithm.
Terminological reasoning is a mode of reasoning all hybrid knowledge representation systems based on KL-ONE rely on. After a short introduction of what terminological reasoning amounts to, it is proven that a complete inference algorithm for the BACK system would be computationally intractable. Interestingly, this result also applies to the KANDOR system, which had been conjectured to realize complete terminological inferences with a tractable algorithm. More generally, together with an earlier paper of Brachman and Levesque it shows that terminological reasoning is intractable for any system using a non-trivial description language. Finally, consequences of this distressing result are briefly discussed. 1 Introduction  The BACK system  1  [13] belongs to the class of hybrid knowledge representation systems based on KL-ONE (cf. the article by Brachman and Schmolze [4]). As in any other system of this family, a frame-based description language (henceforth FDL), which can be viewed as a ...
Incremental Induction of Rules for Document Image Understanding.
This paper aims at presenting the application of first-order logic machine learning techniques to two document domains in order to earn rules fo reco) izing the semanticro eo f theirol:P co: o ents. Specifica y, the mu tistrategy incrementa earning system INTHELEX has been app ied to mu ti-fo rmat scientific papers and do cuments co ncerning Eur o ean fi msfro the 20&#039;s and 30&#039;s. The cha enge co es fro the di#erent eve so f fo rmatting standards in thesedo mains: fro m(mo re o r ess) standardized ayo uts, in scientific papers, to do cuments with a - mo) no standard, in histoR3l cu tura heritage material. Experimenta resu ts in bo th do mains and a c o pariso with the Progol system assess the advantages that the expoitation of INTHELEX can yield. 
A Methodology for the Induction of Ontological Knowledge from Semantic Annotations.

Automatic Induction of Rules for Classification and Interpretation of Cultural Heritage Material.

Evaluation and Validation of Two Approaches to User Profiling.
In the Internet era, huge amounts of data are available to  everybody, in every place and at any moment. Searching for relevant  information can be overwhelming, thus contributing to the user&#039;s sense  of information overload. Building systems for assisting users in this task  is often complicated by the di#culty in articulating user interests in a  structured form - a profile - to be used for searching. Machine learning  methods o#er a promising approach to solve this problem. Our research  focuses on supervised methods for learning user profiles which are predictively  accurate and comprehensible.
Improving Automatic Labelling through RDF Management.

Correcting the Document Layout A Machine Learning Approach.
In this paper, a machine learning approach to support the user during the correction of the layout analysis is proposed. Layout analysis is the process of extracting a hierarchical structure describing the layout of a page. In our approach, the layout analysis is performed in two steps: firstly, the global analysis determines possible areas containing paragraphs, sections, columns, figures and tables, and secondly, the local analysis groups together blocks that possibly fall within the same area. The result of the local analysis process strongly depends on the quality of the results of the first step. We investigate the possibility of supporting the user during the correction of the results of the global analysis. This is done by allowing the user to correct the results of the
Learning Logic Programs for Layout Analysis Correction.
Layout analysis is the process of extracting a  hierarchical structure describing the layout of a  page. In the system WISDOM++, the layout  analysis is performed in two steps: firstly, the  global analysis determines possible areas  containing paragraphs, sections, columns, figures  and tables, and secondly, the local analysis  groups together blocks that possibly fall within  the same area. The result of the local analysis  process strongly depends on the quality of the  results of the first step. We investigate the  possibility of supporting the user during the  correction of the results of the global analysis.
Automated Classification of Web Documents into a Hierarchy of Categories.

An Exhaustive Matching Procedure for the Improvement of Learning Efficiency.

Theta%2DSubsumption and Resolution A New Algorithm.

Personalization to improve searching in a Digital Library.

Mining Model Trees with Regression and Splitting Nodes.

RDF Core A component for effective management of RDF Models.
In order to make Semantic Web effective, the first step was the development of languages that could support data portability, namely XML, metadata descriptions, namely RDF, and ontology management and inference, such as DAML+OIL, OWL etc. Those languages have to be manipulated by applications and many Application Programming Interfaces (APIs) have been developed in order to accomplish this task. Obviously, they differ in implementation details. Moreover, developers often would like to exploit more than an API at a time. Another issue is that a developer would be very advantaged if he could have a uniform support for some services across these frameworks (such as query languages), despite the lack of standards. In this paper, we present a component, called RDFCore, developed in order to overcome these problems. We will also illustrate the added value that our framework provides to RDF in order to exploit the full potentiality of the language and to employ it in research as well as in real world applications. Consequently we will provide some test results on the performances of the presented framework. 1
Incremental Multistrategy Learning for Document Processing.

Empowering a GIS with inductive learning capabilities the case of INGENS.
Informationgfor intopog)A2(U map captions or in GIS models is often insu#cient to recog737 interesting gerestingU patterns. Some prototypes of GIS have already been extended with a knowledg--8U(A and somereasoning capabilities to support sophisticated map interpretation processes. Nevertheless, the acquisition of the necessary knowledg is still a demanding task for which machinelearning techniques can be ofgA(7 help. This paper presents INGENS, a prototypical GIS whichinteg757U machinelearning tools to assist users in the task oftopogA273) map interpretation. The system can be trained to learn operational definitions of gC35UA27(3( objects that are not explicitly modeled in the database. INGENS has been applied to the task of Apulian map interpretation in order to discovergcoverA(6 knowledg of interest to town planners.
An introduction to symbolic data analysis and the SODAS software.

Mining HTML Pages to Support Document Sharing in a Cooperative System.
In this paper, the problem of classifying HTML documents is  investigated in the context of a client-server application, named WebClass,  developed to support the search activity of a geographically distributed group of  people with common interests. The two main issues studied in the paper are the  selection of some features to represent HTML documents and the construction  of the classifiers. A new feature selection technique is presented and its  interaction with different classifiers is experimentally studied. Results show that  performance improves even with simple classifiers and the proposed feature  selection technique compares favorably with respect to other well-known  approaches.
Multistrategy Learning of Rules for Automated Classification of Cultural Heritage Material.
This work presents the application of a new, enhanced version  of theincremfi tal  learningsystem  INTHELEX (INcrem5 tal THEory  Learnerfrom  EXamr5xC: the  learningcom  onent in the architecture  of the EU project COLLATE, dealing with the annotation of cultural  heritage documC ts. Due to thecomx:C shape of the handledmndled5#  the addition  ofm  ultistrategy capabilities was needed toim#: ve the  e#ectiveness and e#ciency of the learning  process.Som resultsdemlt  strating the benefits that the addition of each strategy can bring are also  reported.
Adaptive Layout Analysis of Document Images.

Cooperation of Multiple Strategies for Automated Learning in Complex Environments.

AI%2AIA 2001 Advances in Artificial Intelligence

Document Classification and Interpretation through the Inference of Logic%2DBased Models.

Multiple Predicate Learning for Document Image Understanding.

A Logical Framework for Frequent Pattern Discovery in Spatial Data.

Automated Discovery of Dependencies Between Logical Components in Document Image Understanding.
Document image understanding denotes the recognition of semantically relevant components in the layout extracted from a document image. This recognition process is based on some visual models, whose manual specification can be a highly demanding task. In order to automatically acquire these models, we propose the application of machine learning techniques. In this paper, problems raised by possible dependencies between concepts to be learned are illustrated and solved with a computational strategy based on the separate-and-parallel-conquer search. The approach is tested on a set of real multi-page documents processed by the system WISDOM++. New results confirm the validity of the proposed strategy and show some limits of the learning system used in this work.
OI%2Dimplication Soundness and Refutation Completeness.

First%2DOrder Rule Induction for the Recognition of Morphological Patterns in Topographic Maps.

Mining Association Rules in Spatial Databases A Logic%2Dbased Computational Method.

Machine Learning in Computer Vision %28Editorial%29.

A Generalization Model Based on OI%2Dimplication for Ideal Theory Refinement.

Transforming paper documents into XML format with WISDOM  .
The transformation of scanned paper documents to a form suitable for an Internet browser is a complex process that requires solutions to several problems. The application of an OCR to some parts of the document image is only one of the problems. In fact, the generation of documents in HTML format is easier when the layout structure of a page has been extracted by means of a document analysis process. The adoption of an XML format is even better, since it can facilitate the retrieval of documents in the Web. Nevertheless, an effective transformation of paper documents into this format requires further processing steps, namely document image classification and understanding. WISDOM++ is a document processing system that operates in five steps: document analysis, document classification, document understanding, text recognition with an OCR, and text transformation into HTML/XML format. The innovative aspects described in the paper are: the preprocessing algorithm, the adaptive page segmen...
Inductive learning from numerical and symbolic data An integrated framework.
Numerical induction from data is one of the statistical data analysis tasks, which uses a tabular model, with almost exclusively numerical features, as data representation formalism. The output representations are different: from functions to probability distributions, from surface equations to tables of indexes. One approach to extend the classical data analysis techniques to symbolic objects is the Symbolic Data Analysis: the input and the output of classical techniques are expressed in a symbolic way, so guaranteeing the comprehensibility of both the observations and the results, while the processing techniques, although appropriately adapted, maintain the efficiency of the classical statistical inferential models. Also in the field of Machine Learning several methods have been proposed to extend some inductive approaches from statistical data analysis to data represented as attribute-value couples. Sometimes these approaches transform ideas and principles coming from numerical induction to handle propositional calculus descriptions, otherwise they combine different techniques in order to treat numericalcontinuous data and algebraic-symbolic data differently. The aim here is to improve the efficiency and to preserve the expressive power of the representations during the learning process, and to save the accuracy and flexibility of the numerical techniques during the recognition phase. This kind of integration is more and more complex when using first-order computational learning models, which are useful for handling object descriptions in structured domains, when not only the properties of objects but also the relations between different objects must be considered. The necessity arises from integrating different computational strategies, different knowledge represe...
Interaction Profiling in Digital Libraries through Learning Tools.

Ideal Theory Refinement under Object Identity.

Induction of Recursive Theories in the Normal ILP Setting Issues and Solutions.
Induction of recursive theories in the normal ILP setting is a complex  task because of the non-monotonicity of the consistency property. In this paper  we propose computational solutions to some relevant issues raised by the  multiple predicate learning problem. A separate-and-parallel-conquer search  strategy is adopted to interleave the learning of clauses supplying predicates  with mutually recursive definitions. A novel generality order to be imposed to  the search space of clauses is investigated in order to cope with recursion in a  more suitable way. The consistency recovery is performed by reformulating the  current theory and by applying a layering technique based on the collapsed  dependency graph. The proposed approach has been implemented in the ILP  system ATRE and tested in the specific context of the document understanding  problem within the WISDOM project. Experimental results are discussed and  future directions are drawn.
Discovering Geographic Knowledge The INGENS System.

Machine Learning for Intelligent Processing of Printed Documents.
. A paper document processing system is an information system component which transforms information on printed or handwritten documents into a computer-revisable form. In intelligent systems for paper document processing this information capture process is based on knowledge of the speci#c layout and logical structures of the documents. This article proposes the application of machine learning techniques to acquire the speci#c knowledge required byan  intelligent document processing system, named WISDOM++, that manages printed documents, such as letters and journals. Knowledge is represented by means of decision trees and #rst-order rules automatically generated from a set of training documents. In particular, an incremental decision tree learning system is applied for the acquisition of decision trees used for the classi- #cation of segmented blocks, while a #rst-order learning system is applied for the induction of rules used for the layout-based classi#cation and understanding of d...
Multistrategy Theory Revision Induction and Abduction in INTHELEX.

Guest Editors%27 Introduction.

A Machine Learning Approach to Web Mining.

Conceptual Change in Learning Naive Physics The Computational Model as a Theory Revision Process.

WISDOM   An Interactive and Adaptive Document Analysis System.
WISDOM++ is a document analysis system whose main design requirements are real-time user interaction and adaptivity. This paper presents the two-phased skew estimation algorithm and the adaptive document block segmentation and classification techniques. An evaluation of the performance of some of these tasks is also conducted according to a benchmarking procedure.
CONCERTO%2C Conceptual Indexing%2C Querying and Retrieval of Digital Documents.

Machine Learning for Intelligent Document Processing The WISDOM System.

A Learning Server for Inducing User Classification Rules in a Digital Library Service.

CONCERTO

Symbolic Learning Techniques in Paper Document Processing.

An Adaptive Visual Environment for Digital Libraries.

Learning Recursive Theories with ATRE.

Interacting With IDL The Adaptive Visual Interface.

Interfacce adattative per WWW il caso di IDL.

Adding Machine Learning and Knowledge Intensive Techniques to a Digital Library Service.
This paper presents IDL, a prototypical digital library service. It integrates machine learning tools and intelligent techniques in order to make e#ective, efficient and economically feasible the process of capturing the information that should be stored and indexed by content in the digital library. In fact, information capture and semantic indexing are critical issues when building a digital library, since they involve complex pattern recognition problems, such as document analysis, classification and understanding. Experimental results show that learning systems can e#ectively and e#ciently solve all these problems.
Handling Continuous Data in Top%2DDown Induction of First%2DOrder Rules.
Al~raet. Handling numerical information is one of the most important research issues for practical applications of first-order learning systems. This paper is concerned with the problem of inducing first-order classification rules frum both numeric and symbolic d__ ~ We propose a spocialiTation operator that dis~-~etizes continuous d~_~ _ during the learning proce ~ The heuristic function used to choose among different discretizations satisfies a property that can be profitably exploited to improve the efficiency of the specialiTmlon operator. The operator has been implemented and tested on the domnnent understanding domain. 1 Background and motivation One of the most important research issues for practical applications of first-order learning systems is the handling of numerical information [7,18]. Indeed, in many applications, continuous data are predominant and should be treated suitably. Image analysis and understsnding is one of the widest application fields where the capability of handling both symbolic and numeric descriptors (that is, attributes and relations) is essential for the successftd application of machine learning techniques. In
Processing Paper Documents with WISDOM.

IDL A Prototypical Intelligent Digital Library Service.
wideavailable standard graphical user interfaces (GUIs), and Thispaperpresents a prototypical digital library service. It multimediatcchnology.Oneofthepeculiaritiesofourproject integrates machine learning tools and techniques in order to liesintheroIethatintelligentagents,namelylearningsystems, makeeflective,efticientandeconomicallyfeasible theprocess can play for ir$ormation capture and semantic indexing. of capturbtg the in$onnation that should be stored and By ir$ormation capture, we mean the task of setting indexed by content in the digital library. Infact, information information items free of the physical medium on which they capture is one of the main bottleneck when building a digital am stored It involves the problem of converting data from a library, since it involves complex pattern recognition paper format into a digital one. This has given raise to the problems, such as document analysis, classification and bii of new research areas dealing with the automated understanding. Experimental results show that learning recognition of the components of documents, ranging from systems can solve &amp;ectively and&amp;iciently all these problems. low-level elements, such as characters (OCR), symbols, text, lines, to high-level ones, such as sections, colmrms, graphics, images.handwriting[15].Severalproblemsareaddressedby 1
Machine Learning   On%2Dline Libraries %3D IDL.

Information Capture and Semantic Indexing of Digital Libraries through Machine Learning Techniques.
wideavailable standard graphical user interfaces (GUIs), and Thispaperpresents a prototypical digital library service. It multimediatcchnology.Oneofthepeculiaritiesofourproject integrates machine learning tools and techniques in order to liesintheroIethatintelligentagents,namelylearningsystems, makeeflective,efticientandeconomicallyfeasible theprocess can play for ir$ormation capture and semantic indexing. of capturbtg the in$onnation that should be stored and By ir$ormation capture, we mean the task of setting indexed by content in the digital library. Infact, information information items free of the physical medium on which they capture is one of the main bottleneck when building a digital am stored It involves the problem of converting data from a library, since it involves complex pattern recognition paper format into a digital one. This has given raise to the problems, such as document analysis, classification and bii of new research areas dealing with the automated understanding. Experimental results show that learning recognition of the components of documents, ranging from systems can solve &amp;ectively and&amp;iciently all these problems. low-level elements, such as characters (OCR), symbols, text, lines, to high-level ones, such as sections, colmrms, graphics, images.handwriting[15].Severalproblemsareaddressedby 1
Knowledge Revision for Document Understanding.

A Logic Framework for the Incremental Inductive Synthesis of Datalog Theories.

Discovering Causal Rules in Relational Databases.

Machine Learning for Map Interpretation An Intelligent Tool for Environmental Planning.

A Comparative Analysis of Methods for Pruning Decision Trees.
In this paper, we address the problem of retrospectively pruning decision trees induced from data, according to a topdown  approach. This problem has received considerable attention in the areas of pattern recognition and machine learning, and  many distinct methods have been proposed in literature. We make a comparative study of six well-known pruning methods with the  aim of understanding their theoretical foundations, their computational complexity, and the strengths and weaknesses of their  formulation. Comments on the characteristics of each method are empirically supported. In particular, a wide experimentation  performed on several data sets leads us to opposite conclusions on the predictive accuracy of simplified trees from some drawn in the  literature. We attribute this divergence to differences in experimental designs. Finally, we prove and make use of a property of the  reduced error pruning method to obtain an objective evaluation of the tendency to overprune/underprune ob...
Locally Finite%2C Proper and Complete Operators for Refining Datalog Programs.

Revision of Logical Theories.
The task of theory revision in Inductive Logic Programming is to correct the current theory when contradictions with new examples arise. Confidence models are used to assess and compare the different alternatives of revising a theory in order to choose the best revision. In this paper, we propose an augmentation of the two--tiered confidence model of MOBAL/KRT that allows to combine different heuristic confidence measures and can be adapted to different learning strategies.  Keywords: theory revision, predicate invention. 1 Introduction  The difference between incremental learning systems that receive their examples one by one and adapt their theories accordingly, and one--shot systems plays an important role in Machine Learning and especially in ILP. Here, systems range between highly user--independent, non--incremental methods like Golem [Mug&amp;Fen90], and incremental approaches trying to support the more general task of model building like MOBAL [Wro93]. Though the formers are often v...
Simplifying Decision Trees by Pruning and Grafting New Results %28Extended Abstract%29.

Downward Refinement of Hierarchical Datalog Theories.

A knowledge%2Dbased approach to the layout analysis.
In this paper, we present a hybrid approach to the problem of the document analysis in which the document image is segmented by means of a top-down technique and then basic blocks are grouped bottom-up in order to form complex layout components. In this latter process, called layout analysis,only generic knowledge on typesetting conventions is exploited. Sucha knowledge isindependentof theparticular class of processed documents and turns out valuable for a wide range of documents. Preliminary results of the layout analysis system LEX (Layout Expert) show the methodological validity of this approach. 1
Ideal Refinement of Datalog Programs.
. Refinement operators are exploited to change in an automated way incorrect clauses of a logic program. In this paper, we present four refinement operators for Datalogprogramsand demonstrate that all of them meet the properties of local finiteness, properness, and completeness (ideality). Such operators are based on the quasi-ordering induced upon a set of clauses by the generalization model of q-subsumption under object identity. This model of generalization, as well as the four refinement operators have been implemented in a system for theory revision that proved effective in the area of electronic document classification. 1. Introduction In a logic framework for the inductive synthesis of logic programs, a fundamental problem is the definition of locally finite, proper, and complete (ideal) refinement operators. Indeed, when the aim is to develop incrementally a logic program, that should be correct with respect to its intended model at the end of the development process, it become...
An Analytic and Empirical Comparison of Two Methods for Discovering Probabilistic Causal Relationships.
Abstract. The discovery of causal relationships from empirical data is an important problem in machine learning. In this paper the attention is focused on the inference o fprobabilis tic causal relationships, for which two different approaches, namely Glymour et al.&#039;s approach based on constraints on correlations and Pearl and Verma&#039;s approach based on conditional independencies, have been proposed. These methods differ both in the kind of constraints they consider while selecting a causal model and in the way they search the model which better fits to the sample data. Preliminary experiments show that they are complementary in several aspects. Moreover, the method of conditional independence can be easily extended to the case in which variables have a nominal or ordinal domain. In this case, symbohc learning algorithms can be exploited in order to derive the causal law from the causal model. 1
Traps and Pitfalls when Learning Logical Definitions from Relations.

Avoiding Non%2DTermination when Learning Logical Programs A Case Study with FOIL and FOCL.

multistrategy learning for document recognition.

Negation as a Specializing Operator.

Decision Tree Pruning as a Search in the State Space.

Machine Learning Techniques for Knowledge Acquisition and Refinement.
this paper has been supported by a number of sources. The primary support has come from the Office of Nval Pesearch, N00014-92-K-0186, with additional support from DARPA, N00014-85-K-0878 and N00014-87-K-0874, and the Alcatel European Support Center&#039;s Artificial Intelligence Group, 85-ESC-5004
Automatically Acquiring Knowledge by Digital Maps in Artificial Intelligence Planning Techniques.

Classification in Noisy Environments Using a Distance Measure Between Structural Symbolic Descriptions.

Panel Evaluating and Changing Representation in Concept Acquisition.

Flexible Matching for Noisy Structural Descriptions.
Uncertainty on data often makes the task of perfectly matching two descriptions quite ineffective. In this case, a flexible matching, measuring the similarity of two descriptions rather than their equality, is more useful. According to the convention of connecting similarity to the most common concept of distance, we present a definition of distance measure, based on a probabilistic interpretation of the matching predicate, which can cope with structural deformations. As the problem of matching two formulas of the FOPL is NP-complete, two methods arc presented in order to cope with complexity: firstly, a branch-and-bound algorithm, and secondly, a heuristic method. These ideas are applied to the problem of recognizing office documents in digital form according to their page layout. 1
A Distance Measure for Decision Making in Uncertain Domains.

