Overview of prior-art cross-lingual information retrieval approaches

Prior-art search in patent data has specific properties, which set it apart from other traditional information retrieval processes. One major issue is that patents are usually described in generic terms in order to avoid narrowing down the scope of the inventions. Given the growing amount of patents in different countries using different languages, prior-art search applications nowadays need to find patent claims across languages. This has prompted the current research efforts into how to tackle cross-lingual patent search issues. In this paper, we review the state-of-the art of approaches for cross-lingual prior-art search. This includes cross-lingual information retrieval approaches in general and issues that prevent them from working well for prior-art search. Furthermore, we give a brief overview of existing cross-lingual prior-art search approaches and discuss whether they are able to overcome the problems that traditional cross-lingual retrieval approaches have in this area. Finally, a critical analysis based on this overview is presented and ideas on how to tackle some open research issues in cross-lingual prior-art search are given.

► Study of cross-lingual information retrieval (CLIR). ► Specific problems of prior art patent searching in CLIR highlighted. ► Detailed review of the many approaches to CLIR. ► Future research areas suggested.