Cross-Language Information Retrieval using Dutch Query Translation

This paper describes an elementary bilingual information retrieval experiment. The experiment takes Dutch topics to retrieve relevant English documents using Microsoft SQL Server version 7.0. In order to cross the language barrier between query and document, the researchers use query translation by means of a machine-readable dictionary. The Dutch run was void of the typical natural language processing techniques such as parsing, stemming, or part of speech tagging. A monolingual run was carried out for comparison purposes. Due to limitations in time, retrieval system, translation method, and test collection, there is only a preliminary analysis of the results. Introduction and problem description Cross-Language Information Retrieval (CLIR) systems enable users to formulate queries in their native language to retrieve documents in foreign languages [1]. In CLIR, retrieval is not restricted to the query language. Rather queries in one language are used to retrieve documents in multiple languages. Because queries and documents in CLIR do not necessarily share the same language, translation is needed before matching can take place. This translation step tends to cause a reduction in crosslanguage retrieval performance as compared to monolingual information retrieval. The literature explores four different translation options: translating queries (e.g. [2], [3]), translating documents [4], [5], translating both queries and documents [6], and cognate matching 1 [7]. The prevailing CLIR approach is query translation. The translation of queries is inherently difficult due to the lack of a one-to-one mapping of a lexical item and its meaning. This creates lexical ambiguity. Further, query translation is complicated by the cultural differences between language communities and the way they lexicalize the world around them. These two translation issues create many different translation problems such as lexical ambiguity, lexical mismatches, and lexical holes. In turn, these and other translation problems result in translation errors which impact CLIR retrieval performance. The Cross-Language Evaluation Forum (CLEF) provides a multilingual test collection to study CLIR using European languages. One of the CLEF tasks is bilingual information retrieval. The aim of the bilingual task is the retrieval of documents in a language different from the topic (query) language. Unlike the multilingual task, only two languages are involved and retrieval results are monolingual. For the bilingual run we used the Dutch topic set (40 topics) to retrieve English documents (Los Angeles Times of 1994 – 113,005 documents, 409,600 KB). We were completely oblivious to CLEF and its deadlines but we happened to hear that CLEF results were due in one week. We immediately signed up and started on our mad rush to get results in on time.