Unsupervised data processing for classifier-based speech translator
Computer Speech and Language
Hi-index | 0.00 |
Sehda's 2-way speech translation system, S-MINDS, interprets between provider and patient in routine medical interactions with very high accuracy. Optimizing the system for new tasks or languages requires very little data. New developments include a hybrid translation approach that allows participants to say complex or out-of-domain utterances, the expansion of hands-free functionality, and the ability to deliver the most urgent expressions instantaneously.