Multilingual Summarization in Practice: The Case of Patent Claims

Abstract. Hardly any other type of textual material is as difficult to read and comprehend as patents. Especially the claims in a patent reveal very complex syntactic constructions which are difficult to process even for native speakers, let alone for foreigners who do not master well the language in which the patent is written. Therefore, multilingual summarization is very attractive to practitioners in the field. We propose a multilingual summarizer that operates at the Deep-Syntactic Structures (DSyntSs) as introduced in the Meaning-Text Theory. Firstly, the original claims are linguistically simplified and analyzed down to DSyntSs. Then, syntactic and discursive summarization criteria are applied to the DSyntSs to remove summary irrelevant DSyntS-branches. The pruned DSyntS are transferred into DSyntSs of the language in which the summary is to be generated. For the generation of the summary from the transferred DSyntSs, we use the full fledged text generator MATE