Results of psychological experiments have shown that humans make assumptions, which are not necessarily valid, that they are influenced by their background knowledge and that they reason non-monotonically. These observations show that classical logic does not seem to be adequate for modeling human reasoning. Instead of assuming that humans do not reason logically at all, we take the view that humans do not reason classical logically. Our goal is to model episodes of human reasoning and for this purpose we investigate the so-called Weak Completion Semantics. The Weak Completion Semantics is a Logic Programming approach and considers the least model of the weak completion of logic programs under the three-valued Łukasiewicz logic. As the Weak Completion Semantics is relatively new and has not yet been extensively investigated, we first motivate why this approach is interesting for modeling human reasoning. After that, we show the formal correspondence to the already established Stable Model Semantics and Well-founded Semantics. Next, we present an extension with an additional context operator, that allows us to express negation as failure. Finally, we propose a contextual abductive reasoning approach, in which the context of observations is relevant. Some properties do not hold anymore under this extension. Besides discussing the well-known psychological experiments Byrne’s suppression task and Wason’s selection task, we investigate an experiment in spatial reasoning, an experiment in syllogistic reasoning and an experiment that examines the belief-bias effect. We show that the results of these experiments can be adequately modeled under the Weak Completion Semantics. A result which stands out here, is the outcome of modeling the syllogistic reasoning experiment, as we have a higher prediction match with the participants’ answers than any of twelve current cognitive theories. We present an abstract evaluation system for conditionals and discuss well-known examples from the literature. We show that in this system, conditionals can be evaluated in various ways and we put up the hypothesis that humans use a particular evaluation strategy, namely that they prefer abduction to revision. We also discuss how relevance plays a role in the evaluation process of conditionals. For this purpose we propose a semantic definition of relevance and justify why this is preferable to a exclusively syntactic definition. Finally, we show that our system is more general than another system, which has recently been presented in the literature. Altogether, this thesis shows one possible path on bridging the gap between Cognitive Science and Computational Logic. We investigated findings from psychological experiments and modeled their results within one formal approach, the Weak Completion Semantics. Furthermore, we proposed a general evaluation system for conditionals, for which we suggest a specific evaluation strategy. Yet, the outcome cannot be seen as the ultimate solution but delivers a starting point for new open questions in both areas.

With the advent of the Semantic Web and Semantic Technologies, ontologies have become one of the most prominent paradigms for knowledge representation and reasoning. The popular ontology language OWL, based on description logics, became a W3C recommendation in 2004 and a standard for modelling ontologies on the Web. In the meantime, many studies and applications using OWL have been reported in research and industrial environments, many of which go beyond Internet usage and employ the power of ontological modelling in other fields such as biology, medicine, software engineering, knowledge management, and cognitive systems. However, recent progress in the field faces a lack of well-structured ontologies with large amounts of instance data due to the fact that engineering such ontologies requires a considerable investment of resources. Nowadays, knowledge bases often provide large volumes of data without sophisticated schemata. Hence, methods for automated schema acquisition and maintenance are sought. Schema acquisition is closely related to solving typical classification problems in machine learning, e.g. the detection of chemical compounds causing cancer. In this work, we investigate both, the underlying machine learning techniques and their application to knowledge acquisition in the Semantic Web. In order to leverage machine-learning approaches for solving these tasks, it is required to develop methods and tools for learning concepts in description logics or, equivalently, class expressions in OWL. In this thesis, it is shown that methods from Inductive Logic Programming (ILP) are applicable to learning in description logic knowledge bases. The results provide foundations for the semi-automatic creation and maintenance of OWL ontologies, in particular in cases when extensional information (i.e. facts, instance data) is abundantly available, while corresponding intensional information (schema) is missing or not expressive enough to allow powerful reasoning over the ontology in a useful way. Such situations often occur when extracting knowledge from different sources, e.g. databases, or in collaborative knowledge engineering scenarios, e.g. using semantic wikis. It can be argued that being able to learn OWL class expressions is a step towards enriching OWL knowledge bases in order to enable powerful reasoning, consistency checking, and improved querying possibilities. In particular, plugins for OWL ontology editors based on learning methods are developed and evaluated in this work. The developed algorithms are not restricted to ontology engineering and can handle other learning problems. Indeed, they lend themselves to generic use in machine learning in the same way as ILP systems do. The main difference, however, is the employed knowledge representation paradigm: ILP traditionally uses logic programs for knowledge representation, whereas this work rests on description logics and OWL. This difference is crucial when considering Semantic Web applications as target use cases, as such applications hinge centrally on the chosen knowledge representation format for knowledge interchange and integration. The work in this thesis can be understood as a broadening of the scope of research and applications of ILP methods. This goal is particularly important since the number of OWL-based systems is already increasing rapidly and can be expected to grow further in the future. The thesis starts by establishing the necessary theoretical basis and continues with the specification of algorithms. It also contains their evaluation and, finally, presents a number of application scenarios. The research contributions of this work are threefold: The first contribution is a complete analysis of desirable properties of refinement operators in description logics. Refinement operators are used to traverse the target search space and are, therefore, a crucial element in many learning algorithms. Their properties (completeness, weak completeness, properness, redundancy, infinity, minimality) indicate whether a refinement operator is suitable for being employed in a learning algorithm. The key research question is which of those properties can be combined. It is shown that there is no ideal, i.e. complete, proper, and finite, refinement operator for expressive description logics, which indicates that learning in description logics is a challenging machine learning task. A number of other new results for different property combinations are also proven. The need for these investigations has already been expressed in several articles prior to this PhD work. The theoretical limitations, which were shown as a result of these investigations, provide clear criteria for the design of refinement operators. In the analysis, as few assumptions as possible were made regarding the used description language. The second contribution is the development of two refinement operators. The first operator supports a wide range of concept constructors and it is shown that it is complete and can be extended to a proper operator. It is the most expressive operator designed for a description language so far. The second operator uses the light-weight language EL and is weakly complete, proper, and finite. It is straightforward to extend it to an ideal operator, if required. It is the first published ideal refinement operator in description logics. While the two operators differ a lot in their technical details, they both use background knowledge efficiently. The third contribution is the actual learning algorithms using the introduced operators. New redundancy elimination and infinity-handling techniques are introduced in these algorithms. According to the evaluation, the algorithms produce very readable solutions, while their accuracy is competitive with the state-of-the-art in machine learning. Several optimisations for achieving scalability of the introduced algorithms are described, including a knowledge base fragment selection approach, a dedicated reasoning procedure, and a stochastic coverage computation approach. The research contributions are evaluated on benchmark problems and in use cases. Standard statistical measurements such as cross validation and significance tests show that the approaches are very competitive. Furthermore, the ontology engineering case study provides evidence that the described algorithms can solve the target problems in practice. A major outcome of the doctoral work is the DL-Learner framework. It provides the source code for all algorithms and examples as open-source and has been incorporated in other projects.

La World Wide Web actual está compuesta principalmente por documentos escritos para su presentación visual para usuarios humanos. Sin embargo, para obtener todo el potencial de la web es necesario que los programas de computadoras o agentes sean capaces de comprender la información presente en la web. En este sentido, la Web Semántica es una visión futura de la web donde la información tiene significado exacto, permitiendo así que las computadoras entiendan y razonen en base a la información hallada en la web. La Web Semántica propone resolver el problema de la asignación de semántica a los recursos web por medio de metadatos cuyo significado es dado a través de definiciones de ontologías, que son formalizaciones del conocimiento de un dominio de aplicación. El estándar del World Wide Web Consortium propone que las ontologías sean definidas en el lenguaje OWL, el cual se halla basado en las Lógicas para la Descripción. A pesar de que las definiciones de ontologías expresadas en Lógicas para la Descripción pueden ser procesadas por razonadores estándar, tales razonadores son incapaces de lidiar con ontologías inconsistentes. Los sistemas argumentativos constituyen una formalización del razonamiento rebatible donde se pone especial enfasis en la noción de argumento. Así, la construcción de argumentos permite que un agente obtenga conclusiones en presencia de información incompleta y potencialmente contradictoria. En particular, la Programación en Lógica Rebatible es un formalismo basado en la argumentación rebatible y la Programación en Lógica. En esta Disertación, la importancia de la definición de ontologías para poder llevar a cabo la realización de la iniciativa de la Web Semántica junto con la presencia de ontologías incompletas y potencialmente contradictorias motivó el desarrollo de un marco de razonamiento con las llamadas -ontologías. Investigaciones previas de otros autores, determinaron que un subconjunto de las Lógicas para la Descripción pueden ser traducidas efectivamente a un conjunto de la Programación en Lógica. Nuestra propuesta involucra asignar semántica a ontologías expresadas en Lógicas para la Descripción por medio de Programas Lógicos Rebatibles para lidiar con definiciones de ontologías inconsistentes en la Web Semántica. Esto es, dada una ontología OWL expresada en el lenguaje OWLDL, es posible construir una ontología DL equivalente expresada en las Lógicas para la Descripción. En el caso en que DL satisfaga ciertas restricciones, esta puede ser expresada como un programa DeLP P. Por lo tanto, dada una consulta acerca de la pertenencia de una instancia a a un cierto concepto C expresada con respecto a OWL, se realiza un análisis dialectico con respecto a P para determinar todas las razones a favor y en contra de la plausibilidad de la afirmación C(a). Por otro lado, la integración de datos es el problema de combinar datos residiendo en diferentes fuentes y el de proveer al usuario con una vista unificada de dichos datos. El problema de diseñar sistemas de integración de datos es particularmente importante en el contexto de aplicaciones en la Web Semántica donde las ontologías son desarrolladas independientemente unas de otras, y por esta razón pueden ser mutuamente inconsistentes. Dada una ontología, nos interesa conocer en que condiciones un individuo es una instancia de un cierto concepto. Como cuando se tienen varias ontologías, los mismos conceptos pueden tener nombres distintos para un mismo significado o aún nombres iguales para significados diferentes, para relacionar los conceptos entre dos ontologías diferentes se utilizaron reglas puente o de articulación. De esta manera, un concepto se corresponde a una vista sobre otros conceptos de otra ontología. Mostramos también bajo que condiciones la propuesta del razonamiento con -ontologías puede ser adaptada a los dos tipos de integración de ontologías global-as-view y local-as-view considerados en la literatura especializada. Además, analizamos las propiedades formales que se desprenden de este acercamiento novedoso al tratamiento de ontologías inconsistentes en la Web Semántica. Los principales resultados obtenidos son que, como la interpretación de -ontologías como Programas Lógicos Rebatibles es realizada a través de una función de transformación que preserva la semántica de las ontologías involucradas, los resultados obtenidos al realizar consultas son sensatos. También, mostramos que el operador presentado es además consistente y significativo. El acercamiento al razonamiento en presencia de ontologías inconsistentes brinda la posibilidad de abordar de una manera ecaz ciertos problemas de aplicacion del ámbito del comercio electrónico, donde el modelo de reglas de negocio puede ser especificado en términos de ontologías. Entonces, la capacidad de razonar frente a ontologías inconsistentes permite abordajes alternativos conceptualmente más claros, ya que es posible automatizar ciertas decisiones de negocios tomadas a la luz de un conjunto de reglas de negocio posiblemente inconsistentes expresadas como una o varias ontologías y tener un sistema capaz de brindar una explicación del porque se arribo a una conclusión determinada. En consecuencia, presentamos entonces una aplicación del razonamiento sobre ontologías inconsistentes por medio de la argumentación rebatible al modelado de formularios en la World Wide Web. La noción de los formularios como una manera de organizar y presentar datos ha sido utilizada desde el comienzo de la World Wide Web. Los formularios Web han evolucionado junto con el desarrollo de nuevos lenguajes de marcado, en los cuales es posible proveer guiones de validación como parte del código del formulario para verificar que el signifiado pretendido del formulario es correcto. Sin embargo, para el diseñador del formulario, parte de este significado pretendido frecuentemente involucra otras características que no son restricciones por sí mismas, sino más bien atributos emergentes del formulario, los cuales brindan conclusiones plausibles en el contexto de información incompleta y potencialmente contradictoria. Como el valor de tales atributos puede cambiar en presencia de nuevo conocimiento, los llamamos atributos rebatibles. Propusimos entonces extender los formularios web para incorporar atributos rebatibles como parte del conocimiento que puede ser codifiado por el diseñador del formulario, por medio de los llamados -formularios; dicho conocimiento puede ser especificado mediante un programa DeLP, y posteriormente, como una ontología expresada en Lógicas para la Descripción.

L'étude structurale de l'interactome cellulaire peut conduire à des découvertes intéressantes sur les bases moléculaires de certaines pathologies. La modélisation par homologie et l'amarrage de protéines ("protein docking") sont deux approches informatiques pour modéliser la structure tri-dimensionnelle (3D) d'une interaction protéine-protéine (PPI). Des études précédentes ont montré que ces deux approches donnent de meilleurs résultats quand des données expérimentales sur les PPIs sont prises en compte. Cependant, les données PPI ne sont souvent pas disponibles sous une forme facilement accessible, et donc ne peuvent pas être re-utilisées par les algorithmes de prédiction. Cette thèse présente une approche systématique fondée sur l'extraction de connaissances pour représenter et manipuler les données PPI disponibles afin de faciliter l'analyse structurale de l'interactome et d'améliorer les algorithmes de prédiction par la prise en compte des données PPI. Les contributions majeures de cette thèse sont de : (1) décrire la conception et la mise en oeuvre d'une base de données intégrée KBDOCK qui regroupe toutes les interactions structurales domaine-domaine (DDI); (2) présenter une nouvelle méthode de classification des DDIs par rapport à leur site de liaison dans l'espace 3D et introduit la notion de site de liaison de famille de domaines protéiques ("domain family binding sites" ou DFBS); (3) proposer une classification structurale (inspirée du système CATH) des DFBSs et présenter une étude étendue sur les régularités d'appariement entre DFBSs en terme de structure secondaire; (4) introduire une approche systématique basée sur le raisonnement à partir de cas pour modéliser les structures 3D des complexes protéiques à partir des DDIs connus. Une interface web (http://kbdock.loria.fr) a été développée pour rendre accessible le système KBDOCK / Understanding how the protein interactome works at a structural level could provide useful insights into the mechanisms of diseases. Comparative homology modelling and ab initio protein docking are two computational methods for modelling the three-dimensional (3D) structures of protein-protein interactions (PPIs). Previous studies have shown that both methods give significantly better predictions when they incorporate experimental PPI information. However, in general, PPI information is often not available in an easily accessible way, and cannot be re-used by 3D PPI modelling algorithms. Hence, there is currently a need to develop a reliable framework to facilitate the reuse of PPI data. This thesis presents a systematic knowledge-based approach for representing, describing and manipulating 3D interactions to study PPIs on a large scale and to facilitate knowledge-based modelling of protein-protein complexes. The main contributions of this thesis are: (1) it describes an integrated database of non-redundant 3D hetero domain interactions; (2) it presents a novel method of describing and clustering DDIs according to the spatial orientations of the binding partners, thus introducing the notion of "domain family-level binding sites" (DFBS); (3) it proposes a structural classification of DFBSs similar to the CATH classification of protein folds, and it presents a study of secondary structure propensities of DFBSs and interaction preferences; (4) it introduces a systematic case-base reasoning approach to model on a large scale the 3D structures of protein complexes from existing structural DDIs. All these contributions have been made publicly available through a web server (http://kbdock.loria.fr)

Η παρούσα διατριβή πραγματεύεται το αντικείμενο της μάθησης από δύο διαφορετικές οπτικές γωνίες: την επαγωγική και την παραγωγική μάθηση. Αρχικά, παρουσιάζονται παραμετροποιήσεις στοχαστικών μεθόδων εξόρυξης γνώσης από δεδομένα υπό τη μορφή τεσσάρων καινοτόμων εξατομικευμένων μοντέλων στήριξης ασθενών που πάσχουν από διαταραχές άγχους. Τα τρία μοντέλα προσανατολίζονται στην ανεύρεση πιθανών συσχετίσεων μεταξύ των περιβαλλοντικών παραμέτρων των ασθενών και του επιπέδου άγχους που αυτοί παρουσιάζουν, ενώ παράλληλα προτείνεται και η χρήση ενός Μπεϋζιανού μοντέλου πρόβλεψης του επιπέδου άγχους που είναι πιθανό να εμφανίσει κάποιος ασθενής δεδομένων ορισμένων τιμών του περιβαλλοντικού του πλαισίου εφαρμογής. Αναφορικά με το χώρο της εξόρυξης γνώσης από κείμενο και του μετασχηματισμού συμβολοσειρών, προτείνεται η εκπαίδευση μοντέλων δέντρων αποφάσεων για την αυτόματη μεταγραφή Ελληνικού κειμένου στην αντίστοιχη φωνητική του αναπαράσταση, πραγματοποιείται η στοχαστική μοντελοποίηση όλων των πιθανών μεταγραφικών νορμών από ορθογραφημένα Ελληνικά σε Greeklish και τέλος παρουσιάζεται ένας καινοτόμος αλγόριθμος που συνδυάζει δύο γνωστά για την ικανοποιητική τους απόδοση μέτρα σύγκρισης ομοιότητας αλφαριθμητικών προκειμένου να επιτευχθεί η αυτόματη λημματοποίηση του κειμένου εισόδου. Επιπρόσθετα, στα πλαίσια της ανάπτυξης συστημάτων που θα διευκολύνουν την ανάκτηση εγγράφων ή πληροφοριών προτείνεται η συνδυαστική χρήση του προαναφερθέντος αλγορίθμου λημματοποίησης παράλληλα με τη χρήση ενός πιθανοτικού δικτύου Bayes στοχεύοντας στην ανάπτυξη ενός εύρωστου και ανταγωνιστικού ως προς τις επιδόσεις συστήματος ανάκτησης πληροφοριών. Τέλος, παρουσιάζονται οι προτάσεις μας που αφορούν στο χώρο της παραγωγικής μάθησης και του ελέγχου ικανοποιησιμότητας λογικών εκφράσεων. Συγκεκριμένα περιλαμβάνουν: i) την ανάλυση και εκτενή παρουσίαση μιας καινοτόμας μαθηματικής μοντελοποίησης με την ονομασία AnaLog (Analytic Tableaux Logic) η οποία δύναται να εκφράσει τη λογική που διέπει τους αναλυτικούς πίνακες για προτασιακούς τύπους σε κανονική διαζευκτική μορφή. Mέσω του λογισμού Analog επιτυγχάνεται η εύρεση των κλειστών κλάδων του πλήρως ανεπτυγμένου δέντρου Smullyan, χωρίς να είναι απαραίτητος ο αναλυτικός σχεδιασμός του δέντρου, και ii) την παράθεση ενός αναλυτικού αλγορίθμου που μπορεί να αξιοποιήσει τον φορμαλισμό AnaLog σε ένα πλαίσιο αριθμητικής διαστημάτων μέσω του οποίου μπορούμε να αποφανθούμε για την ικανοποιησιμότητα συμβατικών διαζευκτικών προτασιακών εκφράσεων. / The present dissertation deals with the problem of learning from two different perspectives, meaning the inferential and the deductive learning. Initially, we present our suggestions regarding the parameterization of stochastic data mining methods in the form of four treatment supportive services for patients suffering from anxiety disorders. Three of these services focus on the discovery of possible associations between the patients’ contextual data whereas the last one aims at predicting the stress level a patient might suffer from, in a given environmental context. Our proposals with regards to the wider area of text mining and string conversion include: i) the employment of decision-tree based models for the automatic conversion of Greek texts into their equivalent CPA format, ii) the stochastic modeling of all the existing transliteration norms for the Greek to Greeklish conversion in the form of a robust transcriber and iii) a novel algorithm that is able to combine two well-known for their satisfactory performance string distance metric models in order to address the problem of automatic word lemmatization. With regards to the development of systems that would facilitate the automatic information retrieval, we propose the employment of the aforementioned lemmatization algorithm in order to reduce the ambiguity posed by the plethora of morphological variations of the processed language along with the parallel use of probabilistic Bayesian Networks aiming at the development of a robust and competitive modern information retrieval system. Finally, our proposals regarding logical deduction and satisfiability checking include: i) a novel mathematical formalism of the analytic tableaux methodology named AnaLog (after the terms Analytic Tableaux Logic) which allows us to efficiently simulate the structure and the properties of a complete clausal tableau given an input CNF formula. Via the AnaLog calculus it is made possible to calculate all the closed branches of the equivalent complete Smullyan tree without imposing the need to fully construct it, and ii) a practical application of the AnaLog calculus within an interval arithmetic framework which is able to decide upon the satisfiability of propositional formulas in CNF format. This framework, apart from constituting an illustrative demonstration of the application of the AnaLog calculus, it may also be employed as an alternative conventional SAT system.

This thesis describes the design, development, and installation of a computer program which compiles unit clauses generated in a Prolog-based environment at Argonne National Laboratories into Warren Abstract Machine (WAM) code. The program enhances the capabilities of the environment by providing rapid unification and subsumption tests for the very significant class of unit clauses. This should improve performance substantially for large programs that generate and use many unit clauses.

Hochdurchsatzverfahren quantifizieren eine Vielzahl zellulärer Komponenten, können aber selten deren Interaktionen beschreiben. Daher wurden in den letzten 20 Jahren verschiedenste Netzwerk-Rekonstruktionsmethoden entwickelt. Insbesondere Perturbationsdaten erlauben dabei Rückschlüsse über funktionelle Mechanismen in der Genregulierung, Signal Transduktion, intra-zellulärer Kommunikation und anderen Prozessen zu ziehen. Dennoch bleibt Netzwerkinferenz ein ungelöstes Problem, weil die meisten Methoden auf ungeeigneten Annahmen basieren und die Identifizierbarkeit von Netzwerkkanten nicht aufklären. Diesbezüglich beschreibt diese Dissertation eine neue Rekonstruktionsmethode, die auf einfachen Annahmen von Perturbationsausbreitung basiert. Damit ist sie in verschiedensten Zusammenhängen anwendbar und übertrifft andere Methoden in Standard-Benchmarks. Für MAPK und PI3K Signalwege in einer Adenokarzinom-Zellline generiert sie plausible Netzwerkhypothesen, die unterschiedliche Sensitivitäten von PI3K-Mutanten gegenüber verschiedener Inhibitoren überzeugend erklären. Weiterhin wird gezeigt, dass sich Netzwerk-Identifizierbarkeit durch ein intuitives Max-Flow Problem beschreiben lässt. Dieses analytische Resultat erlaubt effektive, identifizierbare Netzwerke zu ermitteln und das experimentelle Design aufwändiger Perturbationsexperimente zu optimieren. Umfangreiche Tests zeigen, dass der Ansatz im Vergleich zu zufällig generierten Perturbationssequenzen die Anzahl der für volle Identifizierbarkeit notwendigen Perturbationen auf unter ein Drittel senkt. Schließlich beschreibt die Dissertation eine mathematische Weiterentwicklung der Modular Response Analysis. Es wird gezeigt, dass sich das Problem als analytisch lösbare orthogonale Regression approximieren lässt. Dies erlaubt eine drastische Reduzierung des nummerischen Aufwands, womit sich deutlich größere Netzwerke rekonstruieren und neueste Hochdurchsatz-Perturbationsdaten auswerten lassen. / 'Omics' technologies provide extensive quantifications of components of biological systems but rarely characterize the interactions between them. To fill this gap, various network reconstruction methods have been developed over the past twenty years. Using perturbation data, these methods can deduce functional mechanisms in gene regulation, signal transduction, intra-cellular communication and many other cellular processes. Nevertheless, this reverse engineering problem remains essentially unsolved because inferred networks are often based on inapt assumptions, lack interpretability as well as a rigorous description of identifiability. To overcome these shortcoming, this thesis first presents a novel inference method which is based on a simple response logic. The underlying assumptions are so mild that the approach is suitable for a wide range of applications while also outperforming existing methods in standard benchmark data sets. For MAPK and PI3K signalling pathways in an adenocarcinoma cell line, it derived plausible network hypotheses, which explain distinct sensitivities of PI3K mutants to targeted inhibitors. Second, an intuitive maximum-flow problem is shown to describe identifiability of network interactions. This analytical result allows to devise identifiable effective network models in underdetermined settings and to optimize the design of costly perturbation experiments. Benchmarked on a database of human pathways, full network identifiability is obtained with less than a third of the perturbations that are needed in random experimental designs. Finally, the thesis presents mathematical advances within Modular Response Analysis (MRA), which is a popular framework to quantify network interaction strengths. It is shown that MRA can be approximated as an analytically solvable total least squares problem. This insight drastically reduces computational complexity, which allows to model much bigger networks and to handle novel large-scale perturbation data.