The aim of distributional semantics is to design computational techniques that can automatically learn the meanings of words from a body of text. The twin challenges are: how do we represent meaning, and how do we learn these representations? The current state of the art is to represent meanings as vectors - but vectors do not correspond to any traditional notion of meaning. In particular, there is no way to talk about 'truth', a crucial concept in logic and formal semantics. In this thesis, I develop a framework for distributional semantics which answers this challenge. The meaning of a word is not represented as a vector, but as a 'function', mapping entities (objects in the world) to probabilities of truth (the probability that the word is true of the entity). Such a function can be interpreted both in the machine learning sense of a classifier, and in the formal semantic sense of a truth-conditional function. This simultaneously allows both the use of machine learning techniques to exploit large datasets, and also the use of formal semantic techniques to manipulate the learnt representations. I define a probabilistic graphical model, which incorporates a probabilistic generalisation of model theory (allowing a strong connection with formal semantics), and which generates semantic dependency graphs (allowing it to be trained on a corpus). This graphical model provides a natural way to model logical inference, semantic composition, and context-dependent meanings, where Bayesian inference plays a crucial role. I demonstrate the feasibility of this approach by training a model on WikiWoods, a parsed version of the English Wikipedia, and evaluating it on three tasks. The results indicate that the model can learn information not captured by vector space models.

Cette thèse s’inscrit dans le cadre du projet Région Aquitaine - INRIA : ITIPY. Ce projet vise à terme l’extraction automatique d’itinéraires à partir de récits de voyage du XIX ème et du début du XX ème siècle. Notre premier travail fut de caractériser le corpus comme échantillon du français, par une étude contrastive d’une part de données quantitatives et d’autre part de la structure des récits de voyage. Nous nous sommes ensuite consacrée à l’étude du temps, et plus particulièrement à l’analyse automatique de la sémantique des temps verbaux du français. Disposant d’un analyseur syntaxique et sémantique à large échelle du français, basé sur les grammaires catégorielles et la sémantique compositionnelle (λ-DRT), notre tâche a été de prendre en compte les temps des verbes pour reconstituer la temporalité des événements et des états, notions regroupées sous le termes d’éventualité. Cette thèse se concentre sur la construction d’un lexique sémantique traitant des temps verbaux du français. Nous proposons une extension et une adaptation d’un système d’opérateurs compositionnels conçu pour les temps du verbe anglais, aux temps et à l’aspect du verbe français du XIX ème siècle à nos jours. Cette formalisation est de facto opérationnelle, car elle est définie en terme d’opérateurs du λ-calcul dont la composition et la réduction, déjà programmées, calculent automatiquement les représentations sémantiques souhaitées, des formules multisortes de la logique d’ordre supérieur. Le passage de l’énoncé comportant une éventualité seule au discours, dont le maillage référentiel est complexe, est discuté et nous concluons par les perspectives qu’ouvre nos travaux pour l’analyse du discours. / This work has been lead in the frame of the ITIPY project which goal was to automatically extract itineraries from travel novels from the XIX th century and from the beginning of the XX th. Our thesis work is close to the text understanding task in the information retrieval field and we aim at building a representation of meaning of linguistic utterances, leaning on the compositionnality principle. More precisely, the itinerary extraction supposes to temporally represent displacement and localization events or states (that we actually call eventualities) of a traveler as far as we understand it through discourse. Working on an automatic parser for syntax (in categorial grammars) and semantics (in λ-DRT), we focused on the building of an semantic lexicon for tense in French. We actually characterized our corpora as a sample of French language, by the means of a quantitative and qualitative analysis as well as a study of the internal structure of this genre. The main contribution of this work deals with tense and aspect semantic processing of the event expressed by tensed verbs, and with its modelling. In this respect, we propose an adaptation and an extension for French from XIX th century to nowadays of a lexicon originally produced to deal with English verbs. This formalisation is operational, for it is defined in λ-calculus which composition and réduction, already implemented, calculate automatically semantics représentations, high order logic formulas. Transition from a single event uterrance to a whole discourse which contains a complex referential network is discussed and allows us to define the limits of this hereby work.

In software engineering, software reuse has been a popular topic since 1968. Nowadays, Software Product Line (SPL) engineering promotes systematic reuse throughout the whole software development process. Within SPL, reusability strongly depends on variability. In this context, variability modelling and management are crucial activities that crosscuts all development stages. Different techniques are used to model variability and one of them is Feature Diagrams (FDs). FD languages are a family of popular modelling languages used to model, and reason on, variability. Since the seminal proposal of a FD language, namely FODA, many extensions have been proposed to improve it. However, the pros and cons of these languages are difficult to evaluate for two main reasons: (1) most of them are informally defined and (2) no well defined criteria were used to justify the extensions made to FODA. As a consequence, variability modelling and management techniques proposed in the literature or used by practitioners may be suboptimal. Globally, this work underlines that the current research on FDs is fragmented and provides principles to remedy this situation. A formal approach is proposed to introduce more rigour in the motivation, definition and comparison of FD languages. Thereby, examining their qualities should be more focused and productive. A formal approach guarantees unambiguity and is a prerequisite to define formal quality criteria and to produce efficient and safe tool automation. A quality analysis is necessary to avoid the proliferation of languages and constructs that are an additional source of misinterpretations and interoperability problems. The creation or selection of a FD language should be driven and motivated by rigourous criteria. Translations from one FD language to another should be defined and carefully studied to avoid interoperability problems. The main contributions of this work are: (1) to use a quality framework to serve as a roadmap to improve the quality of FD languages, (2) to formally evaluate and compare FD language qualities according to well-defined criteria and following a clear method, (3) to formally define and motivate a new FD language that obtains the best scoring according to the quality criteria and (4) to develop tool support for this language.

In the last 50 years, a considerable amount of research workhas been completed in the mathematical system theory and theoryof control. Implementation of the results from this researchinto practice has drastically decreased the production costs.Most production processes are highly automated, and the use ofrobots in industry is growing. As far as office, or businessprocesses are concerned, the situation is quite different.Though the office workers and sales personnel have obtainedmuch help from the modern computers, the office and salesprocesses are far behind the production processes on the levelof automation. The computers are used in the office mainly tohelp in performing various activities, e.g., to write a letter,to print an invoice, to complete a transaction, etc. Thecontrol of the business processes in the office remains, to alarge extent, manual. There is a lot to gain if the controlover business processes could be automated, at leastpartially. The material presented in this thesis is aimed to supportthe following hypothesis: "The ideas worked out in the Mathematical system theory formodeling and controlling physical processes can be successfullyused for modeling and controlling business processes." Oneof the main ideas of mathematical system theory is to considera process as a set of valid trajectories in a state space, andthis idea is the keystone for the thesis. The thesis startswith reformulating the state-oriented approach for the domainof business processes to show what kind of sate space can beused in this domain. First, the approach is introducedinformally by means of an example. Next, a possibleformalization adjusted to the properties of business processesis discussed. Then, experimental evidences that the methodsuggested in the thesis can be used in practice are presented.The suggested method is also compared with other methods ofbusiness process modeling to find out the areas where it hasadvantages over the other methods. In the conclusion, theresults are summarized, and plansfor the future are drawn. Most of the material included in the thesis has beenpublished and presented at international conferences. Thecontribution of this thesis consists in organizing the materialin support of the main hypothesis.

<p>In the last 50 years, a considerable amount of research workhas been completed in the mathematical system theory and theoryof control. Implementation of the results from this researchinto practice has drastically decreased the production costs.Most production processes are highly automated, and the use ofrobots in industry is growing. As far as office, or businessprocesses are concerned, the situation is quite different.Though the office workers and sales personnel have obtainedmuch help from the modern computers, the office and salesprocesses are far behind the production processes on the levelof automation. The computers are used in the office mainly tohelp in performing various activities, e.g., to write a letter,to print an invoice, to complete a transaction, etc. Thecontrol of the business processes in the office remains, to alarge extent, manual. There is a lot to gain if the controlover business processes could be automated, at leastpartially.</p><p>The material presented in this thesis is aimed to supportthe following hypothesis: "<i>The ideas worked out in the Mathematical system theory formodeling and controlling physical processes can be successfullyused for modeling and controlling business processes</i>." Oneof the main ideas of mathematical system theory is to considera process as a set of valid trajectories in a state space, andthis idea is the keystone for the thesis. The thesis startswith reformulating the state-oriented approach for the domainof business processes to show what kind of sate space can beused in this domain. First, the approach is introducedinformally by means of an example. Next, a possibleformalization adjusted to the properties of business processesis discussed. Then, experimental evidences that the methodsuggested in the thesis can be used in practice are presented.The suggested method is also compared with other methods ofbusiness process modeling to find out the areas where it hasadvantages over the other methods. In the conclusion, theresults are summarized, and plansfor the future are drawn.</p><p>Most of the material included in the thesis has beenpublished and presented at international conferences. Thecontribution of this thesis consists in organizing the materialin support of the main hypothesis.</p>

No description available.

Workflow systems aim to provide automated support for the conduct of certain business processes. Workflow systems are driven by workflow specifications which among others, capture the execution interdependencies between various activities. These interdependencies are modelled by means of different control flow constructors, e.g., sequence, choice, parallelism and synchronisation. It has been shown in the research on workflow patterns that the support for and the interpretation of various control flow constructs varies substantially across workflow systems. Two of the most problematic patterns relate to the OR-join and to cancellation. An OR-join is used in situations when we need to model " wait and see" behaviour for synchronisation. Different approaches assign a different (often only intuitive) semantics to this type of join, though they do share the common theme that synchronisation is only to be performed for active paths. Depending on context assumptions this behaviour may be relatively easy to deal with, though in general its semantics is complicated, both from a definition point of view (in terms of formally capturing a desired intuitive semantics) and from a computational point of view (how does one determine whether an OR-join is enabled?). Many systems and languages struggle with the semantics and implementation of the OR-join because its non-local semantics require a synchronisation depending on an analysis of future execution paths. This may require some non-trivial reasoning. The presence of cancellation features and other OR-joins in a workflow further complicates the formal semantics of the OR-join. The cancellation feature is commonly used to model external events that can change the behaviour of a running workflow. It can be used to either disable activities in certain parts of a workflow or to stop currently running activities. Even though it is possible to cancel activities in workflow systems using some sort of abort function, many workflow systems do not provide direct support for this feature in the workflow language. Sometimes, cancellation affects only a selected part of a workflow and other activities can continue after performing a cancellation action. As cancellation occurs naturally in business scenarios, comprehensive support in a workflow language is desirable. We take on the challenge of providing formal semantics, verification techniques as well as an implementation for workflows with those features. This thesis addresses three interrelated issues for workflows with cancellation regions and OR-joins. The concept of the OR-join is examined in detail in the context of the workflow language YAWL, a powerful workflow language designed to support a collection of workflow patterns and inspired by Petri nets. The OR-join semantics has been redesigned to represent a general, formal, and decidable approach for workflows in the presence of cancellation regions and other OR-joins. This approach exploits a link that is proposed between YAWL and reset nets, a variant of Petri nets with a special type of arc that can remove all tokens from a place. Next, we explore verification techniques for workflows with cancellation regions and OR-joins. Four structural properties have been identified and a verification approach that exploits coverability and reachability notions from reset nets has been proposed. The work on verification techniques has highlighted potential problems with calculating state spaces for large workflows. Applying reduction rules before carrying out verification can decrease the size of the problem by cutting down the size of the workflow that needs to be examined while preserving some essential properties. Therefore, we have extended the work on verification by proposing reduction rules for reset nets and for YAWL nets with and without OR-joins. The proposed OR-join semantics as well as the proposed verification approach have been implemented in the YAWL environment.

[pt] A teoria da verdade de Tarski, também conhecida como teoria clássica da verdade, forneceu uma definição para o conceito de verdade que pode ser considerada adequada do ponto de vista material, no sentido de implicar logicamente todas as definições parciais de verdade para sentenças isoladas, do tipo ´s é verdadeira se e somente se p`, sendo p uma sentença qualquer de uma das linguagens às quais a definição tarskiana de verdade se aplica, e sendo s um nome para tal sentença. Essa definição de verdade é consistente - até onde se sabe - mas para garantir a consistência da mesma Tarski precisou restringir seu campo de aplicação ao conjunto das linguagens que não podem ser utilizadas para tratar de sua própria semântica, às quais chamamos linguagens semanticamente abertas. Uma outra teoria da verdade, devida a J. Barwise e J. Etchemendy, foi desenvolvida com vistas a poder ser aplicada de modo consistente às linguagens semanticamente fechadas. Entretanto, para garantir a consistência de sua teoria, Barwise e Etchemendy acabaram por formulá-la de um modo em que a mesma não pudesse ser considerada materialmente adequada, no sentido especificado acima. Isso poderia causar a impressão de que a consistência de uma definição de verdade exigisse que se abrisse mão da adequação material da mesma, tal como fazem Barwise e Etchemendy, ou então que se restringisse seu campo de aplicação às linguagens semanticamente abertas, tal como faz Tarski. Construindo uma definição de verdade materialmente adequada no sentido em questão, e ao mesmo tempo aplicável a linguagens semanticamente fechadas, contudo, S. Kripke mostrou que isso não é o caso. Após uma análise comparativa dessas três teorias da verdade, encontramos razões para formular uma outra definição de verdade, baseada nas intuições russellianas acerca dessa noção, e capaz de ser aplicada consistentemente às linguagens semanticamente fechadas, mantendo a adequação material no sentido acima descrito. No presente trabalho, podem ser encontradas essa análise comparativa das teorias da verdade de Tarski, Barwise-Etchemendy e Kripke, bem como a definição de verdade que formulamos de modo a satisfazer as condições que mencionamos acima. / [en] The tarskian theory of truth, that is also known as the classical theory of truth, has provided a truth definition that may be considered adequate from a material standpoint. By a materially adequate truth definition we understand a truth definition that logically entails every partial truth definition for a specific sentence, of the kind ´s is true if and only if p`, where p is any sentence from some language that is object of the tarskian truth definition, and where s is a name for such sentence. This truth definition is consistent - so far as we know - but in order to guarantee its consistency Tarski had to restrict its field to the set of the languages which cannot be used to speak about their own semantics. These languages are called semantically open languages. Another truth definition, due to J. Barwise and J. Etchemendy, was developed in order to be consistently applied to semantically closed languages. However, in order to assure the consistency of their theory, Barwise and Etchemendy developed it in such a way that it cannot be considered as a materially adequate theory in the sense that has been specified above. We may conclude from this that in order to guarantee the consistency of a truth definition we have either to abandon its material adequacy or to restrict its field to the set of semantically open languages. Kripke showed that this is not the case through developing a truth definition that is both materially adequate and applicable to semantically closed languages. After doing a comparison between these three truth theories, we have found some reasons to formulate another truth definition, which is based on russellian intuitions concerning the notion of truth, and which is able to be consistently applied to semantically closed languages and is also materially adequate in the sense specified above. In this dissertation one can find the comparison between the truth theories of Tarski, Barwise and Etchemendy, and Kripke, which we just mentioned, and also the truth definition that we formulated in order to satisfy the conditions which we described above.

本論文以兩篇結論對立的社論為個案，藉由形式語義學設計出將論說性篇章刻畫成一階謂詞邏輯符號的論述分析程序，使自然語言中對真實宣稱或正當宣稱的爭論還原成真值語義和論證形式有效性的判斷，從而具體地挽救被歧見所破壞的溝通行動。設計過程初擬了處理語篇命題結構、複合語句成分、回指關係、量號轄域歧義、因果語句、隱性前提推論、和引述結構的實用方案。 / By the means of formal semantics, this thesis used two opposite editorials as cases to design a discourse analysis procedure which translates expository-argumentative texts into formulae of first-order predicate logic. The procedure reduces disputes in natual language about truth or rightness claims into judgements on truth values and validities of argument forms, thus the communicative actions disabled by quarrels can be practically redempted. The designing drafted feasible treatments for propositional structures of discourses, compound sentential constituents, anaphoras, quantifier scope ambiguities, cause-and-effect statements, inferences for absent premises, and quotation structures.

Cette étude vise à unifier dans une même approche les descriptions données par une sémantique cognitive et les représentations associées, et étudiées par une sémantique formelle. La sémantique cognitive a pour but d’associer des schèmes aux significations des unités analysées. La sémantique formelle consiste à étudier les modes de représentation de ces schèmes et les relations avec les observables. Notre étude s’appuie et s’insère dans le modèle général développé à l’université Paris-Sorbonne dans le groupe LaLIC (Langages, Logiques, Informatique et Cognition), il s’agit d’utiliser le modèle de la GAC (Grammaire Applicative et Cognitive) et de la GRACE (GRammaire Applicative Cognitive et Enonciative) ; ces deux modèles font appel d’une part à la topologie et d’autre part à la logique combinatoire en vue d’un traitement automatique de la signification. Nous avons choisi d’étudier le problème des prépositions et plus spécialement des trois prépositions : dans, sous et à du français et de leur équivalents en arabe ; cela nous à conduit à rechercher des invariants associés à ces trois prépositions ou relateurs, la préposition dans renvoie à l’intériorité d’un lieu, qu’il soit spatial, temporel, spatio-temporel, notionnel ou d’activité ; la préposition sous renvoie à un lieu déterminé ou engendré par un autre lieu dont on prend la fermeture ; quand à la préposition à renvoie à la fermeture d’un lieu. Le lieu est un lieu cognitif abstrait, suffisamment général qui, selon le contexte peut prendre des valeurs plus particulières. / This study aims to unify in a single approach the descriptions given by cognitive semantics and associated representations, studied by formal semantics. Cognitive semantics consists to associate the meanings of the analyzed with schemes. Formal semantics consists in studying the modes of representation of these schemes and the relations to the observables. Our study is based on the general model developed at the Paris-Sorbonne University in the LaLIC group, using the GAC model (Applicative and Cognitive Grammar) and GRACE (GRammar Applicative Cognitive and Enunciative), these two models use the one hand to the topology and on the other hand to the combinatory logic in order to an automatic processing of meanings. We have chosen to study the problem of the three prepositions: dans, sous and à of French and their equivalences in Arabic, this leads us to search for invariants associated with these three prepositions or relators, the preposition dans refers to the interiority of a place, be it spatial, temporal, spatial-temporal, notional or activity and the preposition sous refers to a specific place or generated by another place whose closing is taken. The preposition à refers to the closing of a place, here the place is cognitive or abstract place, sufficiently general that according to the context can take more particular values.