Interactive translation of conversational speech
The aim of the paper is to describe JANUS-II, a large scale system devoted to interactive spoken language translation. JANUS-II operates now on spontaneous conversational human-to-human dialogs in limited domains with vocabularies of around 3000+ words. At present, the system accepts English, German, or Spanish input and delivers translations into German, English, Spanish, Japanese or Korean. Further languages are under development. Application of statistical, contextual, prosodic and discourse constraints permits a progressively narrowing search for the most plausible interpretation of an utterance. During translation, JANUS-II produces translations that are used for interactive correction of translation errors. Along with the efforts to improve JANUS-II robustness and accuracy, several system prototypes have been implemented to explore translation needs in different settings: speech translation in one-on-one video conferencing, a portable mobile interpreter, or as passive simultaneous conversation translator. The usability and performance of these systems are discussed. For the entire collection see [Zbl 0927.00044].
- Incremental speech translation.
- The spoken language translator
- scientific article; zbMATH DE number 1949730 (Why is no real title available?)
- scientific article; zbMATH DE number 1949732 (Why is no real title available?)
- scientific article; zbMATH DE number 2080371 (Why is no real title available?)
- scientific article; zbMATH DE number 5197839 (Why is no real title available?)
- The JANUS-III translation system: Speech-to-speech translation in multiple domains
- Nine issues in speech translation
This page was built for publication: Interactive translation of conversational speech
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4942761)