From Documents to Structured Data: First Milestones of the Garzoni Project

Led by an interdisciplinary consortium, the Garzoni project undertakes the study of apprenticeship, work and society in early modern Venice by focusing on a specific archival source, namely the Accordi dei Garzoni from the Venetian State Archives. The project revolves around two main phases with, in the first instance, the design and the development of tools to extract and render information contained in the documents (according to Semantic Web standards) and, as a second step, the examination of such information. This paper outlines the main progress and achievements during the first year of the project.

Exploitation de connaissances sémantiques externes dans les représentations vectorielles en recherche documentaire

Florian Seydoux

The work presented in this thesis deals with several problems met in information retrieval (IR), task which one can summarise as identifying, in a collection of "documents", a subset of documents carrying a sought information, i.e.. relevant for a request expressed by a user. In the case of textual documents, to which we limited ourselves within the framework of this thesis, a significant part of the difficulty lies in ambiguity inherent to human languages. The interaction with the user is also approached in our work, by studying a tool enabling a natural language access to a database. Finally, some techniques which permit the visualisation of large collections of documents are also presented. In this document we first of all describe the principal models of IR by highlighting the relations which exist with some manual technics of IR and document retrieval, developed during the past centuries. We present the principle of document indexing, allowing us to represent documents in a multidimensional space, and the use of this representation by a vectorial model. After having reviewed the principal improvements made these last years with vectorial research systems, including the preprocessings of collections, the indexing mechanism and measurements of similarities between documents, we detail some recent usecases of additional semantic resources (semantic dictionaries, thesaurus, networks, ontologies) reported in scientific literature for the indexing task. We then present more in detail the semantic indexing principle of textual documents by using a thesaurus, consisting in integrating in the document's representation space at least part of the informational contents of hierarchical semantic resources. We propose a general framework allowing us to describe and position various possible techniques to carry out the semantic indexing by adapting, if possible, the specificity of the descriptions resulting from the semantic resources to the data to be represented. We use this framework to describe three families of criteria usable for semantic indexing, each one having its own characteristics. For each of these families, we give the specific algorithms allowing the computation of the criteria. The first two families allow us to consider several criteria already known in feature selection. Moreover we show that, unfortunately, many of these criteria are in fact not very effective for the considered task. The third family allows us to introduce a completely new criterion, the Minimum Redundancy Cut criterion (MRC), built on the basis of the information theory and allowing us to obtain index terms having a probability of occurrence in the collection of documents as well balanced as possible. Finally, we treat the case of semantic index independent of the data (statically choosen), allowing a parameterisation of the level of generality of the index terms. Some of the criteria suggested for semantic indexing has been empirically evaluated. To judge their relevance, we used a well known vectorial system (the Smart IR system) and measured the performances of IR obtained with various reference collections. Those collections was indexed on the basis of the studied criterion, by taking into account the strongly structuring semantic relation of hyper/hyponymy ("is-a" relation), given by two different semantic resources. By comparing results obtained with the performances of a traditional indexing (using the lemmas of the words as representation space), we can show on one hand the relevance of the semantic indexings (in RD) and on the other hand the quality of the proposed criterion (MRC). Concerning man-machine interaction, we present a general outline allowing to build in a relatively fast and systematic way systems with mixed initiative, giving the human user a large (and natural) latitude in the control of the dialogue. This outline is usable in typical database research-task applications (where the database is hidden to the user, but the latter knows exactly which information they wish to find) as well as advice-task applications, for which the users does not necessarily have a precise idea of their needs, and uses the system not only for specifing their wishes, but also a set of propositions as a final result. We particularly stress the techniques allowing us to obtain a robust system, able to deal with speech recognizer failures. Concerning the visualisation of large textual data collections, we present an application of the correspondences analysis (allowing to highlight similarities and oppositions for various groups of entity, built on the basis of additional features present in the DB) to the case of patents data. In addition, we propose a method (based on the bootstrap replication principle) allowing us to determine a confidence interval for relative positionings of various groups, thus permit to immediately judge the reliability of the visually apparent similarities or oppositions.

EPFL2006

Analysis of information processing in the olfactory bulb by in vivo experiments and theoretical modelling

Brice Bathellier

Much information is conveyed to animals by the diverse molecules propagated in their environment. This information is transmitted and treated within the olfactory system to be utilized by the rest of the brain. In this process, the sensory flow arriving from the nose first encounters the olfactory bulb. This thesis studies different aspects of the encoding and treatment of olfactory information in the bulb by a combination of experimental and theoretical approaches. The olfactory bulb consists of two functional stages: an input stage where the nerve terminals from the sensory neurons of the nose are spatially arranged in a precise manner, and an output stage corresponding to the neurons which propagate the processed information towards the more central brain areas. The first part of this thesis is focused on the development of an image processing technique to precisely map the input stage. In a second part, output neurons activity in responsive areas of the bulb was recorded in anaesthetized mice. The data obtained indicate complex dynamics in neural activity on several time scales. We show that the combination of mean firing rates from a large enough ensemble of neurons allows to accurately predict the odor presented to the animal. We also show that temporal variables related to the dynamics give very little information complementary to the ensemble rate code, suggesting that these variables are not read by brain structures downstream to the bulb. In a third part, we developed a numerical model of the olfactory bulb to understand its fast oscillatory dynamics. The model proposes a mechanism based on the negative feedback from the interneurons that regulate output neurons. Some predictions of the model are checked experimentally. An analytical interpretation of the model is also proposed and its generalization to the analysis of fast oscillation in other parts of the brain is discussed.

EPFL2007

Integrated, web-based information system to support collaborative resource management

François Golay, Jens Ingensand

In order to support a more effective and environmentally aware resource management, information systems may be designed to integrate relevant information and indicators about the resource and its environment and to make them available to all stakeholders and producers. Moreover, the development of such an information system towards an open, web-based platform supporting not only information retrieval and visualization, but also data collection and sharing, fosters collaborative resource management processes. As an exemple of such an integrated information system supporting knowledge sharing, the paper presents the RIV project, aiming to develop an interactive system for wine-cultivation based on open-source software. For wine-cultivation it is essential to have knowledge about environmental resources (such as water, soils, micro-climate), but also about growing-techniques and other factors. A first step was therefore to calculate and to collect the information of all wine-growing regions in a spatial database. One main idea behind RIV was to spread this information to all actors involved in wine-cultivation business through the Internet. Another important idea behind RIV was to offer a parcel-management system that integrates with the data that has been collected. The users thus add their own data (both spatial- and non-spatial) and save costs. In order to develop RIV we integrated all main actors in the project and in the planning and development process.