Passage retrieval and intellectual property in legal texts
Santiago Correa, Davide Buscaldi, Paolo Rosso, Alfonso Ríos · 2009
Question Answering (QA) can be viewed as a particular form of Information Retrieval (IR), in which the amount of information to return is the minimum required to satisfy the user needs expressed by a specific question such as: "Where is the Europol Drugs Unit?”1. A Passage Retrieval (PR) system is an IR system which, given a list of keywords (e.g.: "Electricity," "Motor", etc..) or a question such as the previous one, returns fragments of texts (passages) that are relevant to the user needs. The Cross-Language Evaluation Forum2 (CLEF), organises competitions for the assessment of multilingual information retrieval systems. In CLEF-2009 edition, due to the growing interest in natural language processing of legal texts from both the university and the business sector, tracks such as ResPubliQA3 and IP4 have been organised. We have participated in both tracks in the framework of the collaboration between the Natural Language Engineering Lab. of the Universidad Politécnica de Valencia (UPV) and the Maat Knowledge enterprise. In order to address both tracks on QA in legal texts and on Intellectual Property (IP) of patent retrieval, we have used the JIRS (JAVA Information Retrieval System) search engine, a freely available5 PR system which has been developed at the UPV [1]. The results have been sent to