公开(公告)号：US09378731B2

公开(公告)日：2016-06-28

申请号：US14693268

申请日：2015-04-22

Applicant: Google Inc.

Inventor： Olga Kapralova ,  John Paul Alex ,  Eugene Weinstein ,  Pedro J. Moreno Mengibar ,  Olivier Siohan ,  Ignacio Lopez Moreno

IPC: G10L15/00 ,  G10L15/06 ,  G10L25/30 ,  G10L15/187 ,  G10L15/26

CPC classification number: G10L15/063 ,  G10L15/01 ,  G10L15/16 ,  G10L15/187 ,  G10L15/30 ,  G10L15/32 ,  G10L2015/0633

Abstract: The present disclosure relates to training a speech recognition system. One example method includes receiving a collection of speech data items, wherein each speech data item corresponds to an utterance that was previously submitted for transcription by a production speech recognizer. The production speech recognizer uses initial production speech recognizer components in generating transcriptions of speech data items. A transcription for each speech data item is generated using an offline speech recognizer, and the offline speech recognizer components are configured to improve speech recognition accuracy in comparison with the initial production speech recognizer components. The updated production speech recognizer components are trained for the production speech recognizer using a selected subset of the transcriptions of the speech data items generated by the offline speech recognizer. An updated production speech recognizer component is provided to the production speech recognizer for use in transcribing subsequently received speech data items.

Abstract translation: 本公开涉及训练语音识别系统。 一个示例性方法包括接收语音数据项集合，其中每个语音数据项对应于先前由生产语音识别器提交用于转录的话语。 生产语音识别器使用初始生产语音识别器组件来产生语音数据项的转录。 使用离线语音识别器生成每个语音数据项的转录，并且将离线语音识别器组件配置为与初始制作语音识别器组件相比提高语音识别精度。 使用由离线语音识别器生成的语音数据项的转录的所选择的子集来对生产语音识别器进行更新的制作语音识别器组件的训练。 更新的生产语音识别器组件被提供给生产语音识别器，用于转录随后接收的语音数据项。

公开(公告)号：US20160180214A1

公开(公告)日：2016-06-23

申请号：US14577301

申请日：2014-12-19

Applicant: Google Inc.

Inventor： Dimitri Kanevsky ,  Ignacio Lopez Moreno ,  Dmitrii Vladimirovich Ulianov

IPC: G06N3/08 ,  G10L15/06 ,  G06N99/00

CPC classification number: G06N3/08 ,  G06N3/0454 ,  G10L15/063 ,  G10L2015/088

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. One of the methods includes training a neural network using sharp discrepancy learning by providing training data to the neural network, calculating a gradient using a sharp discrepancy output layer objective function to classify the neural network parameters for correct and incorrect network model states, and training the neural network using the gradient to determine a probability that data received by the neural network has features similar to key features of one or more keywords or key phrases.

Abstract translation: 方法，系统和装置，包括在计算机存储介质上编码的用于训练神经网络的计算机程序。 其中一种方法包括通过向神经网络提供训练数据来训练使用尖锐差异学习的神经网络，使用尖锐差异输出层目标函数计算梯度，以将神经网络参数分类为正确和不正确的网络模型状态，并训练 神经网络使用梯度来确定由神经网络接收的数据的概率具有与一个或多个关键词或关键短语的关键特征相似的特征。