XisQuê XisQuê (beta version) is a question answering service for the Web of documents written in the Portuguese language. It handles domain-indepedent factual questions. It is based on a prototype developed, mantained and being expanded by the NLX—Natural Language and Speech Group at the University of Lisbon, Department of Informatics. |
||
Features XisQuê takes a factual question phrased in Portuguese as input and tries to find a short and exact answer for it in the documents written in Portuguese available on the Web. In its current online version, XisQuê shows up to five possible answers ordered by decreasing plausibility. For each answer, the service also returns the sentence where the answer was found (termed long-answer), together with a link to the original document. The system currently supports the following four types of factual questions:
|
||
Performance The system's performance was assessed over a test-set of 60 questions —15 for each type of question handled by the system— randomly selected from cards of the Trivial Pursuit® game. These questions can be found in the annex below. The following results were obtained with the March, 2008 version of the system: Recall: A short-answer was returned for 57% of the questions. Some answer (short- or long-) was returned for 98% of the questions. Precision: Considering the 5 tentative answers provided for each question, a correct answer (short or long) was found to 98% of the questions in the test-set. For 55% of the question, a correct short-answer was found. For further details, see this paper. |
||
Authorship XisQuê is being developed by António Branco, Lino Rodrigues, João Silva and Sara Silveira, with the contribution of Mariana Avelãs and Carolina Silva (MultiWordnet) at the NLX—Natural Language and Speech Group at the University of Lisbon, Department of Informatics. |
||
Acknowledgments The development of XisQuê was partly supported by the FCT— Foundation for Science and Technology of the MCT— Portuguese Ministery of Science and Technology, in the project QueXting, under the contract POSI/PLP/61490/2004, in the scope of the POS_Conhecimento program. |
||
White Papers Branco, António, Lino Rodrigues, João Silva and Sara Silveira, 2008, "Real-time Open-Domain QA in the Portuguese Web". LNAI 5290, Springer. pp. 322-331. Branco, António, Lino Rodrigues, João Silva and Sara Silveira, 2008, "XisQuê: An Online QA Service for Portuguese", In Proceedings of the International Conference on the Computational Processing of Portuguese (PROPOR2008), Berlin, Springer. Avelãs, Mariana, António Branco, Rosa del Gaudio and Carolina Silva, subm., "Projecting a Portuguese Ontology by Triangulation to Support Open-Domain Question-Answering". Ferreira, Eduardo, João Balsa and António Branco, 2007, "Combining Rule-based and Statistical Methods for Named Entity Recognition in Portuguese", Anais do XXVII Congresso da Sociedade Brasileira de Computação, pp.1615-1624, TIL2007—V Workshop em Tecnologia da Informação e da Linguagem Humana. Rodrigues, Lino, 2007, Infra-estrutura de um Serviço Online de Resposta-a-Perguntas com base na Web Portuguesa, Dissertação de Mestrado, Departamento de Informática da Faculdade de Ciências da Universidade de Lisboa. Branco, António, Francisco Costa and Filipe Nunes, 2007, "Processing Verb Inflection Ambiguity: Toward a characterization of the problem space", Actas do XXII Encontro Anual da Associação Portuguesa de Linguística, Faculdade de Letras de Coimbra. Silva, João, 2007, Shallow Processing of Portuguese: From Sentence Chunking to Nominal Lemmatization, Dissertação de Mestrado, Departamento de Informática da Faculdade de Ciências da Universidade de Lisboa. Branco, António and João Silva, 2007, "Very High Accuracy Rule-based Nominal Lemmatization with a Minimal Lexicon", In Actas do XXII Encontro Anual da Associação Portuguesa de Linguística, Faculdade de Letras de Coimbra. Branco, António and Francisco Costa, 2007, "Identification and Handling of Dialectal Variation with a Single Grammar", In: Peter Dirix, Ineke Schuurman, Vincent Vandeghinste, and Frank Van Eynde (eds.) Proceedings of the 17th Meeting of Computational Linguistics in the Netherlands (CLIN17), Utrect, LOT, pp.5-19. Barreto, Florbela, António Branco, Eduardo Ferreira, Amália Mendes, Maria Fernanda Nascimento, Filipe Nunes and João Silva, 2006a, "Open Resources and Tools for the Shallow Processing of Portuguese", Proceedings of the 5th International Conference on Language Resources and Evaluation (LREC2006), pp. 1438-1443. Barreto, Florbela, António Branco, Eduardo Ferreira, Amália Mendes, Maria Fernanda Nascimento, Filipe Nunes and João Silva, 2006b, "Linguistic Resources and Software for Shallow Processing", Actas do XXI Encontro Anual da Associação Portuguesa de Linguística, pp.203-218. Branco, António and João Silva, 2006a, "Dedicated Nominal Featurization of Portuguese". Lecture Notes in Artificial Intelligence 3960, Berlim, Springer, ISSN03029743, pp.244-247. Branco, António and Francisco Costa, 2006, "Noun Ellipsis without Empty Categories", Proceedings of the 13th International Conference on Head-Driven Phrase Structure Grammar, Stanford, CSLI Publications, pp.81-101. Branco, António and João Silva, 2006b, "LX-Suite: Shallow Processing Tools for Portuguese", Proceedings of the 11th Conference of the European Chapter of the Association for Computational Linguistics (EACL2006), Trento, Itália, pp.179-182. Rodrigues, Lino, 2006, Relatório Parcial de Implementação do Sistema QueXting, Dez. 2006, Departamento de Informática da Faculdade de Ciências da Universidade de Lisboa. |
||
Contact Us Even if not for donations ;-), contact us through nlxgroup@NOSPAM (replace "NOSPAM" by "di.fc.ul.pt"). |
||
Annex — Test-set
|