Modeling and improving text stability in live captions

Automatic speech recognition (ASR) technology has made conversations more accessible with live captions in remote conferencing software, mobile applications, and head-worn displays. However, to maintain real-time responsiveness, live caption systems often display interim predictions that are updated as new utterances are received. This can cause text instability (a “flicker” where previously displayed text is updated, shown in the captions on the left in the …

applications asr augmented reality automatic speech recognition captions conferencing conversations engineer google hci head mobile mobile applications modeling natural-language understanding predictions reality real-time recognition research research scientist software software engineer speech speech recognition stability systems technology text