Catálogo de publicaciones - libros

We present a keyphrase extraction algorithm for scientific publications. Different from previous work, we introduce features that capture the positions of phrases in document with respect to logical sections found in scientific discourse. We also introduce features that capture salient morphological phenomena found in scientific keyphrases, such as whether a candidate keyphrase is an acronyms or uses specific terminologically productive suffixes. We have implemented these features on top of a baseline feature set used by Kea [1]. In our evaluation using a corpus of 120 scientific publications multiply annotated for keyphrases, our system significantly outperformed Kea at the  < .05 level. As we know of no other existing multiply annotated keyphrase document collections, we have also made our evaluation corpus publicly available. We hope that this contribution will spur future comparative research.

This paper describes our efforts to develop a toolset and process for automated metadata extraction from large, diverse, and evolving document collections. A number of federal agencies, universities, laboratories, and companies are placing their collections online and making them searchable via metadata fields such as author, title, and publishing organization. Manually creating metadata for a large collection is an extremely time-consuming task, but is difficult to automate, particularly for collections consisting of documents with diverse layout and structure. Our automated process enables many more documents to be available online than would otherwise have been possible due to time and cost constraints. We describe our architecture and implementation and illustrate the effectiveness of the tool-set by providing experimental results on two major collections DTIC (Defense Technical Information Center) and NASA (National Aeronautics and Space Administration).

Syllabi are important documents created by instructors for students. Gathering syllabi that are freely available, and creating useful services on top of the collection, will yield a digital library of value for the educational community. However, gathering and building a repository of syllabi is complicated by the unstructured nature of syllabus representation and the lack of a unified vocabulary for syllabus construction. In this paper, we propose an intelligent approach to automatically annotate freely-available syllabi from the Web to benefit the educational community through supporting services such as semantic search. We discuss our detailed process for converting unstructured syllabi to structured representations through entity recognition, segmentation, and association. Our evaluation results demonstrate the effectiveness of our extractor and also suggest improvements. We hope our work will benefit not only users of our services but also people who are interested in building other genre-specific repositories.

In today’s knowledge-intensive engineering environment, information management is an important and essential activity. However, existing researches of Engineering Information Management (EIM) mainly focused on numerical data such as computer models and process data. Textual data, especially the case of free texts, which constitute a significant part of engineering information, have been somewhat ignored, mainly due to their lack of structure and the noisy information contained in them. Since summarization is a process to distill important information from source documents and at the same time remove irrelevant and redundant information, it could address the obstacles for handling textual data in EIM. Moreover, text summarization could address the increasing demand to integrate information from multiple documents and reduce the time in acquiring useful information from massive textual data in the engineering domain. This paper discusses in detail the need to apply text summarization in EIM and introduces a case study in summarizing multiple online customer reviews.

This paper describes the PENG project that integrates personalized push and pull technologies to access relevant information. PENG integrates several key tasks, including personalized filtering, retrieval, and presentation of multimedia news, into a single system. In this paper we provide an overview of PENG, describing our approach to constructing a dedicated retrieval and content management system for a specific user group. We also report critically on the results of a user and task based evaluation.

This paper proposes a system for realizing a new children’s reading environment in a library. Conventional libraries establish children’s rooms for supporting children’s activities of reading printed picture books. The collections there, however, are often biased in content and limited in number, and children are usually not satisfied with them. When the entire collection becomes large, it occurs that younger children cannot effectively search that collection. Similar search problems also exist in children’s digital libraries. We have designed a system called BrowsReader with the purpose of realizing a new environment, where children can browse in a large virtual bookshelf and can easily find the digitized and/or printed picture books that they are interested in. A user study was conducted to test the effectiveness of the system.

This work investigates the potential of applying a suite of visualisation for query processing in a desktop search environment. While each of these visualizations may not be totally new on its own, we have attempted to add value to each one by endowing it with useful features, and to seamless integrate them to allow easy switching of views, thereby providing the novelty in this work to create a potentially useful means to process search results and carry out query refinements and exploration. These visualisations include a List View, Tree View, Map View, Bubble View, Tile View and Cloud View. A first evaluation was undertaken by 94 M.Sc. participants to gauge the system’s potential usefulness and to detect usability issues with its interface and graphical presentations. The evaluation results were encouraging and showed that these views to be both effective and useful on the whole, and support the research premise that a combination of integrated visualisations will result in a more effective search tool.

DIGMAP is a project to find solutions for digital libraries scenarios focused on digitised historical maps. The main service will reuse metadata from European national libraries and other relevant third party metadata sources to provide discovery and access to contents. This will also include a proof of concept of a scenario of reusing and enriching these metadata by automatic processes that will try to extract relevant indexing information from the images of the digitised maps, as well as from any kind of associated text.

SMS or short messaging service is a form of text messaging used extensively throughout the world. It is a cheap and mobile form of communication but there is limited research into its library use. Using Internet search techniques and content analysis, this research investigated how libraries use SMS. Fifty libraries with English language websites were found to be using SMS and were divided evenly between academic and public, with two national libraries. They spread over fourteen countries with the United Kingdom having the most libraries using the technology. Usage was mainly in the circulation area, particularly reserves. Some libraries offered reference services. Cost and complexity varied, with some libraries offering examples of possible future use. Suggestions are made as to the importance of libraries helping each other in implementing SMS.

Topic detection and analysis is very important to understand academic document collections. By further modeling the influence among the topics, we can understand the evolution of research topics better. This problem has attracted much attention recently. Different from the existing works, this paper proposes a solution which discovers hidden topics as well as the relative change of their intensity as a first step and then uses them to construct a module network. Through this way, we can produce a generalization module among different topics. In order to eliminate the instability of topic intensity for analyzing topic changes, we adopt the piece-wise linear representation so that we can model the topic influence accurately. Some experiments on real data sets validate the effectiveness of our proposed method.