Holopedia is a cooperative project involving two research groups:
Natural Language Processing & Information Retrieval Group (NLP&IR-UNED) at UNED group has had as a central research topic (since its creation in 1995) the application of advanced Natural Language Processing Techniques to Intelligent Information Access. Teh group has collaborated with other relevant Spanish research groups from UPC and UB in Barcelona, UPV/EHU in the Basque Country, University of Alicante, U. Politécnica de Valencia, U. Jaén, CSIC, in projects funded by the Spanish Government such as ITEM (TIC96-1243-C03-01), Hermes (TIC2000-0335-C03-01), R2D2 (TIC2003-07158-C04), Text-Mess (TIN2006-15265-C06-02) and QEAVIs (TIN2007-67581-C02-01)).
NLP&IR-UNED has also collaborated with relevant European research players from academia and industry in EC-funded projects such as EuroWordNet (4FP, LE2-4003), CLEF (5FP, ST-2000-31002) and TrebleCLEF (7FP, ICT-1-4-1, grant #215231), MultiMatch (6FP, STREP 033104) and MedIEQ (6FP, Public Health 61383). Of particular interest to this proposal is WePS, an international evaluation campaign for systems that search for information about people and organizations in the Web, that is run by UNED in cooperation with New York University, University of Chicago at Illinois, Intelius Corp. (a U.S. company providing Web People Search services) and Llorente & Cuenca (one of the leading communication companies in Spain).
The group has a long and successful profile in developing advanced Information Access technologies, and has participated in several scientific competitions such as Word Sense Disambiguation at Senseval, Cross-Language Text and Image Retrieval at CLEF, Question Answering and Answer Validation at CLEF-QA, and Web People Search at WePS.
The Barcelona Media Group (BM) combines both a wide expertise in Information Retrieval (MinWeb project, especially in ranking, web analysis, semantic web) and in Natural Language Processing. The group has experience in developing NLP components (Named Entity Classification components, dependency parsingand Time-aware annotation in multilingual contexts as well as using high-throughput text analytics to elicit fine-grained knowledge from user-generated content that is frequently noisy and ambiguous). The researchers involved in this proposal involved have a strong publication record applying and developing these technologies in fields as diverse as Semantic Search, Business Intelligence and Bioinformatics. The acquired knowledge through the participation on the Living Knowledge project on dealing with temporal references and bias will be the base for the proposed task on predicting the future. The group’s expertise in information extraction and Opinion Mining task (integra project) dovetails adequately with the NLP-based processing needed to generate the features required to perform effective document clustering using semantic vectors.
The group keeps also a close collaboration with Yahoo! Labs Barcelona on the semantic annotation of text (SW1, EnglishWikipedia Semantically annotated) and the implementation and design of new paradigm search application which exploits Natural Language Techniques (Correlator, Quest) which are publically deployed in Yahoo!Sandbox. The group has also a background on automatic generation and management of intelligent multimedia content (i3media and Buscamedia projects). The group also participates in projects funded by the national government (minWeb, i3media, Buscamedia, integra) and the European Union (LivingKnowledge) cooperating with other groups from universities and companies (UPC, University of Trento, University of Pavia, University of Southampton, SORA, Activa-Multimedia, and Yahoo! Labs Barcelona).