Файл:Garbage Modeling for On-device Speech Recognition.pdf

Материал из Материалы по машинному обучению
Перейти к: навигация, поиск
Garbage_Modeling_for_On-device_Speech_Recognition.pdf(0 × 0 пикселей, размер файла: 292 КБ, MIME-тип: application/pdf)

Christophe Van Gysel1, Leonid Velikovich2, Ian McGraw3, Franc¸oise Beaufays3 1University of Amsterdam, Amsterdam, The Netherlands 2Google, Inc., New York, NY USA 3Google, Inc., Mountain View, CA USA cvangysel@uva.nl, leonidv@google.com, imcgraw@google.com, fsb@google.co


User interactions with mobile devices increasingly depend on voice as a primary input modality. Due to the disadvantages of sending audio across potentially spotty network connections for speech recognition, in recent years there has been growing attention to performing recognition on-device. The limited computational resources, however, typically require additional model constraints. In this work, we explore the task of on-device utterance verification, wherein the recognizer must transcribe an utterance if it is in a target set or reject it as being out of domain. We present a data-driven methodology for mining tens of thousands of target phrases from an existing corpus. We then compare two common garbage-modeling approaches to utterance verification: a sub-word rejection model and a white-listed n-gram model. We examine a deficiency of the sub-word modeling approach and introduce a novel modification that makes use of common prefixes between targeted phrases and non-targeted phrases. We show good performance inthetrade-offbetweenrecallandworderrorrateusingboththe prefixandwhite-listedn-gramapproaches. Finally,weevaluate the prefix-based approach in a hybrid setting where rejected instances are sent to a server-side recognizer.

Index Terms: automatic speech recognition, language modeling, utterance verification, OOV rejection, garbage modeling

История файла

Нажмите на дату/время, чтобы просмотреть, как тогда выглядел файл.

текущий17:13, 22 декабря 20160 × 0 (292 КБ)Slikos (обсуждение | вклад)
  • Вы не можете перезаписать этот файл.

Следующая 1 страница ссылается на данный файл: