公开(公告)号：US20180025720A1

公开(公告)日：2018-01-25

申请号：US15615799

申请日：2017-06-06

Applicant: Intel Corporation

Inventor： Joachim HOFER ,  Georg STEMMER

IPC: G10L15/02 ,  G10L15/14

CPC classification number: G10L15/02 ,  G10L15/142 ,  G10L2015/025 ,  G10L2015/0635

Abstract: A method in a computing device for decoding a weighted finite state transducer (WFST) for automatic speech recognition is described. The method includes sorting a set of one or more WFST arcs based on their arc weight in ascending order. The method further includes iterating through each arc in the sorted set of arcs according to the ascending order until the score of the generated token corresponding to an arc exceeds a score threshold. The method further includes discarding any remaining arcs in the set of arcs that have yet to be considered.

公开(公告)号：US20160379629A1

公开(公告)日：2016-12-29

申请号：US14750185

申请日：2015-06-25

Applicant: INTEL CORPORATION

Inventor： Joachim HOFER ,  Georg STEMMER ,  Josef BAUER

IPC: G10L15/22 ,  G10L15/193 ,  G06F17/27 ,  G10L15/26 ,  G10L15/24 ,  G10L15/187 ,  G10L15/18

CPC classification number: G06F17/274 ,  G06F17/277 ,  G06F17/2775 ,  G10L15/083 ,  G10L15/19

Abstract: A system, article, and method of automatic speech recognition with dynamic vocabularies is described herein.

Abstract translation: 本文描述了具有动态词汇表的自动语音识别的系统，文章和方法。

公开(公告)号：US20170323638A1

公开(公告)日：2017-11-09

申请号：US15525893

申请日：2014-12-12

Applicant: INTEL CORPORATION

Inventor： Lukasz MALINOWSKI ,  Piotr MAJCHER ,  Georg STEMMER ,  Piotr ROZEN ,  Joachim HOFER ,  Josef BAUER

IPC: G10L15/193 ,  G10L15/22

CPC classification number: G10L15/193 ,  G10L15/083 ,  G10L15/22 ,  G10L15/34

Abstract: A system, article, and method of automatic speech recognition using parallel processing for weighted finite state transducer-based speech decoding.

公开(公告)号：US20160093292A1

公开(公告)日：2016-03-31

申请号：US14499049

申请日：2014-09-26

Applicant: Intel Corporation

Inventor： Joachim HOFER ,  Georg STEMMER

IPC: G10L15/02 ,  G10L15/14

CPC classification number: G10L15/02 ,  G10L15/142 ,  G10L2015/025 ,  G10L2015/0635

Abstract: A method in a computing device for decoding a weighted finite state transducer (WFST) for automatic speech recognition is described. The method includes sorting a set of one or more WFST arcs based on their arc weight in ascending order. The method further includes iterating through each arc in the sorted set of arcs according to the ascending order until the score of the generated token corresponding to an arc exceeds a score threshold. The method further includes discarding any remaining arcs in the set of arcs that have yet to be considered.

Abstract translation: 描述了用于解码用于自动语音识别的加权有限状态换能器（WFST）的计算装置中的方法。 该方法包括按照升序排列一组或多个基于其弧权重的WFST弧。 该方法还包括根据升序迭代排序的弧组合中的每个弧，直到与弧相对应的所生成的令牌的得分超过分数阈值。 该方法还包括丢弃尚未考虑的弧组中的任何剩余弧。