Catálogo de publicaciones - libros

For thousands of years, libraries have allowed humanity to collect and organize data and information, and to support the discovery and communication of knowledge, across time and space. Coming together in this Internet Age, the world’s societies have extended this process to span from the personal to the global, as the concepts, practices, systems, and services related to Library and Information Science unfold through digital libraries. Scientists, scholars, teachers, learners, and practitioners of all kinds benefit from the distributed and collaborative knowledge environments that are at the heart of the digital library movement. Digital libraries thus encompass the dimensions in the 5S Framework: Societies, Scenarios, Spaces, Streams, and Structures. To clarify this approach, we explain the role of meta-models, such as of a minimal digital library (DL), and of more specialized (discipline-oriented) DLs, such as archeological DLs. We illustrate how suitable knowledge environments can be more easily prepared as instances of these meta-models, resulting in usable and useful DLs, including for education, computing, and archaeology.

The article discusses changes in scientific work (academic and applied) associated with new potentials, but also coercions of information technologies. Background for this interest is the experience gained in several digital library projects that inclinations and willingness to use these technical possibilities is much less common than the developers of these systems, and we all, tended to think in recent years. This seems to be true even in those scientific disciplines which were and are at the forefront of the development, e.g. physics, mathematics, etc. The background for this observation is discussed looking at general economic and social changes, viewing the environments of work in the scientific sphere, the contents and their quantity and quality of supply in scientific IT systems, the user side in their communities of practice, and the technological and organizational basis of scientific information. Some strategic issues to improve the situation are discussed in the final part of the paper.

In contrast to standard digital libraries, systems addressing the specific requirements of cultural heritage need to deal with digitized material like scanned documents instead of borne digital items. Such systems aim at providing the means for domain experts, e.g. historians, to collaboratively work with the given material. To support their work, automatic indexing mechanisms for both textual and pictorial digitized documents need to be combined with retrieval methods exploiting the content as well as the context of information items for precise searches. In the COLLATE project we devised several access methods using textual contents, feature extraction from images, metadata, and annotations provided by the users.

The benefits of separating document structure from presentation have long been understood—e.g., the distinction between the SGML and DSSSL standards as well as the ODA standard’s separation of layout and logical structure. More recently, the breadth of applications incorporating XML specifications have provided further evidence in the context of the World-Wide Web of the strength of the abstraction that this separation provides. The structure/presentation separation, focused on describing the characteristics of documents, usefully can be extended to encompass the additional characteristics of interactive hypertextual documents, such as the Web’s—for example, specifying the hypertext’s responses to the reader’s actions remains outside of the scope of the structure/presentation representation. We have explored one such family of models in which the hypertext is modeled by an automaton structure rather than a graph structure. In this paper, I will discuss how these new articulation points have lead to investigations into novel and flexible hypertext/hypermedia system implementations.

Search engines are among the most important services on the Web. Due to the scale of the ever-growing Web, classic centralized models and algorithms can no longer meet the requirements of a search system for the whole Web. Decentralization seems to be an attractive alternative. Consequently Web retrieval has received growing attention in the area of peer-to-peer systems. Decentralization of Web retrieval methods, in particular of text-based retrieval and link-based ranking as used in standard Web search engines have become subject of intensive research. This allows both to distribute the computational effort for more scalable solutions and to share different interpretations of the Web content to support personalized and context-dependent search. In this paper we first review existing studies about the algorithmic feasibility of realizing peer-to-peer Web search using text and link-based retrieval methods. From our perspective realizing peer-to-peer Web retrieval also requires a common framework that enables interoperability of peers using different peer-to-peer search methods. Therefore in the second part we introduce a common framework consisting of an architecture for peer-to-peer information retrieval and a logical framework for distributed ranking computation.

With the increasing importance of the “Global Information Society“ and as the world’s depositories of online collections proliferate, there is a growing need for systems that enable access to information of interest wherever and however it is stored, regardless of form or language. In recognition of this, five years ago, the DELOS Network for Digital Libraries launched the Cross-Language Evaluation Forum (CLEF), with the objective of promoting multi lingual information access by providing the research community with an infra structure for testing and evaluating systems operating in multilingual contexts and a common platform for the comparison of methodologies and results. In this paper, we outline the various activities initiated by CLEF over the years in order to meet the emerging needs of the application communities, and trace the impact of these activities on advances in multilingual system development.

As more information becomes available electronically, tools for finding information of interest to users become increasingly important. Information preferences vary greatly across users, therefore, filtering systems must be highly personalized to serve the individual interests of the user. Our research deals with learning approaches to build user profiles that accurately capture user interests from content (documents) and that could be used for personalized information filtering. The learning mechanisms analyzed in this paper are relevance feedback and a naïve Bayes method. Experiments conducted in the context of a content-based profiling system for movies show the pros and cons of each method.

In information retrieval, feedback provided by individual users is often very sparse. Consequently, machine learning algorithms for automatically retrieving documents or recommending items may not achieve satisfactory levels of accuracy. However, if one views users as members of a larger user community, then it should be possible to leverage similarities between different users to overcome the sparseness problem. The paper proposes a collaborative machine learning framework to exploit inter-user similarities. More specifically, we present a kernel-based learning architecture that generalizes the well-known Support Vector Machine learning approach by enriching content descriptors with inter-user correlations.

In an age characterized by tremendous technological breakthroughs, the world is witnessing overwhelming quantities and types of information. Digital Libraries (DLs) are a result of these breakthroughs, but they have not been spared by the challenges resulting from them. While DLs stakeholders are still struggling to come to terms with the massive quantities and complex types of information, the needs of the digital library as an information/knowledge environment is still evolving including new challenging needs. Information Visualization (Infovis) represents a viable solution to this. The human-vision channel has a high bandwidth and it can surveil a visual field in a parallel manner, processing the corresponding data to different levels of detail and recognition and understanding of overwhelming data can be done at an instant. It is an outstanding resource that can be exploited within a DLs in order to address issues arising from the conventional needs (such as the quantities and types of information) and the non-conventional needs. Here we analyze Infovis as a resource for DLs, relating visualization techniques to specific DLs needs, providing a classification of Infovis techniques and reporting about our analysis of DLs tasks and their correspondence with suitable visualizations.

Research on information visualization has so far established an outline of the information visualization process and shed light on a broad range of detail aspects involved. However, there is no model in place that describes the nature of information visualization in a coherent, detailed, and well-defined way. We believe that the lack of such a lingua franca hinders communication on and application of information visualization techniques. Our approach is to design a declarative language for describing and defining information visualization techniques. The information visualization modelling language (IVML) provides a means to formally express, note, preserve, and communicate structure, appearance, behaviour, and functionality of information visualization techniques and applications in a standardized way. The anticipated benefits comprise both application and theory.