The universal adoption of the Internet and the WWW have created an enormous, multilingual virtual textual database. Rather than looking upon foreign language documents as distracting noise, one can consider these documents as untapped sources of information. This book addresses the problem of accessing multilingual information through a single-language query, a research problem which is receiving growing attention by US and foreign governments. It describes the problem, highlighting the differences between the field and the related areas of machine translation and information retrieval. Researchers from Europe, Japan and America present a wide variety of techniques and experimental results. The life-size experiments are run on modern large-scale retrieval testbeds, running up to hundreds of megabytes of texts. The techniques involve using bilingual dictionaries, machine translation systems, parallel text corpora, comparable but non-parallel text corpora, latent semantic indexing, and weighted Boolean interrogation.
This text should be suitable as a secondary text for a graduate level course on cross-language information retrieval, and as a reference for researchers and practitioners in industry.