No description available.

Indiana University-Purdue University Indianapolis (IUPUI) / After the initial network is constructed using expert’s knowledge of the domain, Bayesian networks need to be updated as and when new data is observed. Literature mining is a very important source of this new data. In this work, we explore what kind of data needs to be extracted with the view to update Bayesian Networks, existing technologies which can be useful in achieving some of the goals and what research is required to accomplish the remaining requirements. This thesis specifically deals with utilizing causal associations and experimental results which can be obtained from literature mining. However, these associations and numerical results cannot be directly integrated with the Bayesian network. The source of the literature and the perceived quality of research needs to be factored into the process of integration, just like a human, reading the literature, would. This thesis presents a general methodology for updating a Bayesian Network with the mined data. This methodology consists of solutions to some of the issues surrounding the task of integrating the causal associations with the Bayesian Network and demonstrates the idea with a semiautomated software system.

No description available.

Advances in microcomputer display devices and support software during the past decade have made the microcomputer an increasingly popular vehicle for technical education. This is especially apparent in the area of simulation. The pedagog can provide the student of control theory not merely with the block diagrams and differential algebra but with high resolution color graphic animations supported by mathematical models whose parameters are easily changed by some editor facility. This mode of control system design and behavior study is both faster and more enjoyable for the student, providing greater continuity, concentration, and learning efficiency. This paper describes the simulation of a PID two-tank level control system. The System is at most fourth-order and provides a good introduction to system control theory. The model is not unusual, and its nonlinear fourth-order Runge-Kutta solution is straightforward. Simulation itself takes form as (1) a graphical animation in which the user is aware of changing water levels and pipe flows and (2) a numerical multi-column output of system inputs, state variables, and outputs. Both applications are very user-friendly. A special editor is developed, under which the above applications run. This paper is not a thorough treatment of the control system; a course is required here. Instead, the focus is on the editor and the organization of its Pascal source code. Discussed are a general editor concept and object-oriented code template to which any mathematical driver and associated simulators may be adapted. The editor’s source code is designed to be programmer-friendly so that the uninitiated programmer may rapidly assimilate editor structure and continue development.

A new type of software, termed a "group editor", allows multiple users to create and simultaneously edit a single document; this software has ostensibly been developed to increase efficiency in co-authoring environments where users may not be co-located. However, questions as to the effectiveness of this type of communication aid, which is a member of the "groupware" family of tools used for some types of computer supported cooperative work, remain. Particularly, there has been very little objective data on any group editor because of the problems inherent in evaluating writing, as well as due to the few examples of group editors that exist. A method was developed to examine the effect of using a particular group editor, Aspects™ from Group Technologies in Arlington, Va., in conjunction with several communication media, on a simple dyad writing task. Six dyads of college students familiar with journalistic writing were matched on attributes of dominance and writing ability and were asked to write short news articles based on short video clips in a balanced two factor within-subject analysis of variance design. Six conditions were tested based on communication media: audio only, audio plus video, and face-to-face; each of these with and without the availability of the group editor. Constraints inherent in the task attempted to enforce consistent document quality levels, measured by grammatical quality and content quality (correctness of information and chronological sequencing). Time to complete the articles was used as a measure of efficiency, independent from quality due to the consistent quality levels of the resulting work. Results from the time data indicated a significant effect of communication media, with the face-to-face conditions taking significantly less time to complete than either of the other media alternatives. Grammatical quality of the written articles was found to be of consistent high quality by way of computerized grammar checker. Content quality of the documents did not significantly differ for any of the conditions. A supplemental Latin square analysis showed additional significant differences in time to complete for trial means (a practice effect) and team differences. Further, significantly less variance was found in certain conditions which had the group editor than in other conditions which did not. Subjective data obtained from questionnaires supported these results and additional1y showed that subjects significantly preferred trials with the group editor and considered then more productive. The face-to-face conditions may have been more efficient due to the nature of the task or due to increased communication structure within dyads due to practice with the group editor. The significant effect of Team Differences may have been due to consistent style differences between dyads that affected efficiency levels. The decreased variability in time to complete in certain group editor conditions may have been due to increased communication structure in these conditions, or perhaps due to leveling effects of group writing as opposed to individual writing with team member aid. These hypotheses need to be tested with further study, and generalizability of the experimental task conditions and results from this particular group editor need to be established as well face-to-face conditions clearly resulted in the most efficient performance on this task. The results obtained concerning the group editor suggest possible efficiency or consistency benefits from the use of group editors by co-authoring persons when face-to-face communication is not practical. Perhaps group editors will become a useful method for surrogate travel for persons with disabilities. / Master of Science

No description available.

TRAC is a text-processing language for use with a reactive typewriter. The thesis describes the design and implementation of a TRAC processor for the Fairchild F24 computer. Chapter I introduces some text processing concepts, the TRAC operations, and the implementation procedures. Chapter II examines the history and -characteristics of the TRAC language. The next chapter specifies the TRAC syntax and primitive functions. Chapter IV covers the algorithms used by the processor. The last chapter discusses the design experience from programming the processor, examines the reactive action caused by the processor, and suggests adding external storage primitive functions for a future version of the processor.

The exponential growth of scientific literature is creating an increased need for systems to process and assimilate knowledge contained within text. Literature Based Discovery (LBD) is a well established field that seeks to synthesize new knowledge from existing literature, but it has remained primarily in the theoretical realm rather than in real-world application. This lack of real-world adoption is due in part to the difficulty of LBD, but also due to several solvable problems present in LBD today. Of these problems, the ones in most critical need of improvement are: (1) the over-generation of knowledge by LBD systems, (2) a lack of meaningful evaluation standards, and (3) the difficulty interpreting LBD output. We address each of these problems by: (1) developing indirect relatedness measures for ranking and filtering LBD hypotheses; (2) developing a representative evaluation dataset and applying meaningful evaluation methods to individual components of LBD; (3) developing an interactive visualization system that allows a user to explore LBD output in its entirety. In addressing these problems, we make several contributions, most importantly: (1) state of the art results for estimating direct semantic relatedness, (2) development of set association measures, (3) development of indirect association measures, (4) development of a standard LBD evaluation dataset, (5) division of LBD into discrete components with well defined evaluation methods, (6) development of automatic functional group discovery, and (7) integration of indirect relatedness measures and automatic functional group discovery into a comprehensive LBD visualization system. Our results inform future development of LBD systems, and contribute to creating more effective LBD systems.

Le traitement automatique des documents écrits est un domaine très actif dans le monde industriel. En effet, devant la masse de documents écrits à traiter, l'analyse automatique devient une nécessité mais les performances des systèmes actuels sont très variables en fonction des types de documents traités. Par exemple, le traitement des documents manuscrits non contraints reste une problématique non encore résolue à ce jour car il existe toujours deux verrous technologiques qui freinent la mise en place de systèmes fiables de traitement automatique des documents manuscrits : - la première concerne la reconnaissance des écritures manuscrites ; - la seconde est liée à l'existence d'une grande variabilité de structures de documents. Cette thèse porte sur la résolution de ce deuxième verrou dans le cas de documents manuscrits non contraints. Pour cela, nous avons développé des méthodes fiables et robustes d'analyse de structures de documents basées sur l'utilisation de Champs Aléatoires Conditionnels. Le choix des Champs Aléatoires Conditionnels est motivé par la capacité de ces modèles graphiques à prendre en compte les relations entre les différentes entités du document (mots, phrases, blocs, ...) et à intégrer des connaissances contextuelles. De plus, l'utilisation d'une modélisation probabiliste douée d'apprentissage permet de s'affranchir de la variabilité inhérente des documents à traiter. L'originalité de la thèse porte également sur la proposition d'une approche hiérarchique permettant l'extraction conjointe des structures physique (segmentation du document en blocs, lignes, ...) et logique (interprétation fonctionnelle de la structure physique) en combinant des caractéristiques physiques de bas niveau (position, représentation graphique, ...) et logiques de haut niveau (détection de mots clés). Les expérimentations effectuées sur des courriers manuscrits montrent que le modèle proposé représente une solution intéressante de par son caractère discriminant et sa capacité naturelle à intégrer et à contextualiser des caractéristiques de différentes natures.

La plupart des travaux en acquisition de lexiques bilingues à partir de corpus comparables reposent sur l'hypothèse distributionnelle qui a été étendue au scénario bilingue. Deux mots ont de fortes chances d'être en relation de traduction s'ils apparaissent dans les mêmes contextes lexicaux. Ce postulat suppose donc une définition claire et rigoureuse du contexte et une connaissance parfaite des indices contextuels. Or, la complexité et les spécificités de chaque langue font qu'il n'est pas aisé d'énoncer une telle définition qui garantisse une extraction de couples de traductions, efficace dans tous les cas de figure. Toute la difficulté réside dans la manière de définir, d'extraire et de comparer ces contextes dans le but de construire des lexiques bilingues fiables. Nous nous efforcerons tout au long des différents chapitres de cette thèse à essayer de mieux comprendre cette notion de contexte, pour ensuite l'étendre et l'adapter afin d'améliorer la qualité des lexiques bilingues. Une première partie des contributions vise à améliorer l'approche directe qui fait office de référence dans la communauté. Nous proposerons plusieurs manières d'aborder le contexte des mots pour mieux les caractériser. Dans la deuxième partie des contributions, nous commencerons par présenter une approche qui vise à améliorer l'approche par similarité inter-langue. Ensuite, une méthode nommée Q-Align, directement inspirée des systèmes de question/réponse sera présentée. Enfin, nous présenterons plusieurs transformations mathématiques et donc plusieurs représentations vectorielles, pour nous concentrer essentiellement sur celles que nous aurons choisi pour développer une nouvelle méthode d'alignement.