Felix Weninger
Transcribing conversations between doctor and patients, as in Nuance’s Dragon Ambient eXperience (DAX) solution, is a very challenging application for speech recognition. One reason is that today’s speech recognition systems can reach optimal performance only for close-talk input, which requires the speaker to wear a headset microphone or use a handheld device like a mobile […]
Deep learning technology has rapidly transformed the way that computers perform speech recognition. It has enabled us to build speech recognizers for very challenging applications such as Dragon Ambient eXperience (DAX), which transcribes conversations between doctor and patient. In particular, the end-to-end (E2E) speech recognition system has been a primary focus of research in recent […]
A crucial capability of automatic speech recognition (ASR) systems is to cope with variability in speech, caused by various accents, age groups, or other variations in speaking style, as well as noisy environments. Several years ago, Nuance’s Dragon dictation product line pioneered the usage of deep learning technology for speaker adaptation in professional dictation systems. […]
