公开(公告)号：US20170220925A1

公开(公告)日：2017-08-03

申请号：US15394617

申请日：2016-12-29

Applicant: Google Inc.

Inventor： Ouais Alsharif ,  Rohit Prakash Prabhavalkar ,  Ian C. McGraw ,  Antoine Jean Bruguier

IPC: G06N3/04 ,  G06N3/08

CPC classification number: G06N3/0445 ,  G06N3/08

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for implementing a compressed recurrent neural network (RNN). One of the systems includes a compressed RNN, the compressed RNN comprising a plurality of recurrent layers, wherein each of the recurrent layers has a respective recurrent weight matrix and a respective inter-layer weight matrix, and wherein at least one of recurrent layers is compressed such that a respective recurrent weight matrix of the compressed layer is defined by a first compressed weight matrix and a projection matrix and a respective inter-layer weight matrix of the compressed layer is defined by a second compressed weight matrix and the projection matrix.

公开(公告)号：US09443517B1

公开(公告)日：2016-09-13

申请号：US14709745

申请日：2015-05-12

Applicant: Google Inc.

Inventor： Jakob Nicolaus Foerster ,  Christopher Walter George Thornton ,  Rohit Prakash Prabhavalkar

IPC: G10L15/16 ,  G10L15/22 ,  G10L15/08 ,  G10L15/06 ,  G10L15/32 ,  G10L25/30

CPC classification number: G10L25/30 ,  G10L15/063 ,  G10L15/08 ,  G10L15/16 ,  G10L25/51 ,  G10L2015/088 ,  G10L2015/223

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. One of the methods includes accessing a first neural network that was trained to recognize a given keyword or keyphrase using a set of hotword training data, wherein the hotword training data includes positive hotword training data that correspond to utterances of the keyword or keyphrase, and negative hotword training data that corresponds to utterances of words or phrases that are other than the keyword or keyphrase, selecting a seed hotsound, mapping, to a feature space, (i) the positive hotword training data, (ii) the negative hotword training data, and (iii) the seed hotsound, performing an optimization of a position of the seed hotsound within the feature space to generate a modified seed hotsound, generating a set of hotsound training data using the modified seed hotsound, training a second neural network to recognize the modified seed hotsound using the generated set of hotsound training data, and using the trained second neural network to recognize the modified hotsound.

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。 其中一种方法包括访问被训练以使用一组热词训练数据来识别给定关键词或关键短语的第一神经网络，其中所述热词训练数据包括对应于关键词或关键短语的话语的正面词汇训练数据，以及否定 对应于除了关键字或关键短语之外的单词或短语的话语的热词训练数据，选择种子hotsound，映射到特征空间，（i）正面词语训练数据，（ii）否定词语训练数据， 并且（iii）种子hotsound，在特征空间内进行种子hotsound的位置的优化以产生修改的种子hotsound，使用修改的种子hotsound生成一组热保护训练数据，训练第二神经网络以识别 使用所生成的一组热刺训练数据来修改种子hotsound，并且使用训练有素的第二神经网络来识别修改后的hotsound。