Speech to Speech Translation: Present and Future Challenges

Significant progress has been made in the field of human language technologies. Various tasks like continuous speech recognition for large vocabulary, speaker and language identification, spoken information inquiry, information extraction and cross-language retrieval in restricted domains are today feasible and different prototypes and systems are running. The spoken translation problem on the other hand is still a significant challenge: “Good text translation was hard enough to pull off. Speech to speech MT was beyond going to the Moon – it was Mars...” [Steve Silbermann, Wired Magazine]. Research issues and approaches to the spoken translation problem will be reviewed by considering present projects and achieved results. Moreover foreseen applications offered by the Web to multilingual person to person communication will be introduced. Issues related to portability, language resources and evaluation will also be discussed. Finally a video of NESPOLE! project, a common EU NSF funded project, exploring future applications in the e-commerce and e-service sectors will be shown.