Browsing Multilingual Information with the MultiSemCor Web Interface

Parallel and comparable corpora represent a crucial resource for different Natural Language Processing tasks like machine translation, lexical acquisition, and knowledge structuring but are also suitable to be consulted by humans for different purposes, such as linguistic teaching, corpus linguistics, translation studies, lexicography, multilingual information browsing. To enhance their exploitation by human users, specially designed interfaces need to be developed. In this paper we present the design and implementation of the MultiSemCor Web Interface. MultiSemCor is a parallel English/Italian corpus, which is being developed at ITC-irst starting from the English corpus SemCor. In MultiSemCor the texts are aligned at word level and semantically annotated with WordNet senses. The MultiSemCor Web Interface allows the users to exploit at best the potentiality of the corpus. We will describe the main functions of the interface, which provides two distinct browsing modalities: a bi-text-oriented modality and a word-oriented modality, which amounts to a bilingual semantic concordancer. Moreover, the MultiSemCor Web Interface is integrated with the on-line MultiWordNet browser, which gives access to the reference lexicon for MultSemCor