Cross-lingual document retrieval categorisation and navigation based on distributed services

The widespread use of the Internet across countries has increased the need for access to document collections that are often written in languages different from a user’s native language. In this paper we describe Clarity, a Cross Language Information Retrieval (CLIR) system for English, Finnish, Swedish, Latvian and Lithuanian. Clarity is a fully-fledged retrieval system that supports the user during the whole process of query formulation, text retrieval and document browsing. We address four of the major aspects of Clarity: (i) the user-driven methodology that formed the basis for the iterative design cycle and framework in the project, (ii) the system architecture that was developed to support the interaction and coordination of Clarity’s distributed services, (iii) the data resources and methods for query translation, and (iv) the support for Baltic languages. Clarity is an example of a distributed CLIR system built with minimal translation resources and, to our knowledge, the only such system that currently supports Baltic languages.