公开(公告)号：US09972310B2

公开(公告)日：2018-05-15

申请号：US14985560

申请日：2015-12-31

Applicant: Interactive Intelligence Group, Inc.

Inventor： Srinath Cheluvaraja ,  Ananth Nagaraja Iyer

IPC: G10L15/16 ,  G10L15/14 ,  G10L25/24 ,  G10L25/27 ,  G06N3/02

CPC classification number: G10L15/144 ,  G06N3/02 ,  G10L15/14 ,  G10L15/16 ,  G10L25/24 ,  G10L25/27

Abstract: A system and method are presented for neural network based feature extraction for acoustic model development. A neural network may be used to extract acoustic features from raw MFCCs or the spectrum, which are then used for training acoustic models for speech recognition systems. Feature extraction may be performed by optimizing a cost function used in linear discriminant analysis. General non-linear functions generated by the neural network are used for feature extraction. The transformation may be performed using a cost function from linear discriminant analysis methods which perform linear operations on the MFCCs and generate lower dimensional features for speech recognition. The extracted acoustic features may then be used for training acoustic models for speech recognition systems.

公开(公告)号：US09489943B2

公开(公告)日：2016-11-08

申请号：US14515607

申请日：2014-10-16

Applicant: Interactive Intelligence Group, Inc.

Inventor： Zhenhao Ge ,  Vivek Tyagi ,  Aravind Ganapathiraju ,  Ananth Nagaraja Iyer ,  Scott Allen Randal ,  Felix Immanuel Wyss

IPC: G10L15/187 ,  G06F17/28 ,  G09B19/04 ,  G10L15/06 ,  G09B19/06

CPC classification number: G10L15/063 ,  G06F17/2735 ,  G06F17/28 ,  G09B19/04 ,  G09B19/06 ,  G10L15/14 ,  G10L15/187 ,  G10L2015/081

Abstract: A system and method for learning alternate pronunciations for speech recognition is disclosed. Alternative name pronunciations may be covered, through pronunciation learning, that have not been previously covered in a general pronunciation dictionary. In an embodiment, the detection of phone-level and syllable-level mispronunciations in words and sentences may be based on acoustic models trained by Hidden Markov Models. Mispronunciations may be detected by comparing the likelihood of the potential state of the targeting pronunciation unit with a pre-determined threshold through a series of tests. It is also within the scope of an embodiment to detect accents.

Abstract translation: 公开了用于学习语音识别的替代发音的系统和方法。 通过发音学习可以覆盖另类名称发音，这些发音先前未被一般发音词典涵盖。 在一个实施例中，在单词和句子中检测电话级和音节级错误可以基于由隐马尔可夫模型训练的声学模型。 可以通过一系列测试来比较目标语音单元的潜在状态与预定阈值的可能性来检测微分。 检测重音也属于实施例的范围。

公开(公告)号：US20150255075A1

公开(公告)日：2015-09-10

申请号：US14638198

申请日：2015-03-04

Applicant: Interactive Intelligence Group, Inc.

Inventor： Srinath Cheluvaraja ,  Ananth Nagaraja Iyer ,  Aravind Ganapathiraju ,  Felix Immanuel Wyss

IPC: G10L19/005

CPC classification number: G10L19/005 ,  G10L15/02 ,  G10L15/08 ,  G10L15/14 ,  G10L15/20 ,  G10L25/18 ,  G10L25/21 ,  G10L2015/025 ,  G10L2019/0012

Abstract: A system and method are presented for the correction of packet loss in audio in automatic speech recognition (ASR) systems. Packet loss correction, as presented herein, occurs at the recognition stage without modifying any of the acoustic models generated during training. The behavior of the ASR engine in the absence of packet loss is thus not altered. To accomplish this, the actual input signal may be rectified, the recognition scores may be normalized to account for signal errors, and a best-estimate method using information from previous frames and acoustic models may be used to replace the noisy signal.

Abstract translation: 提出了一种用于在自动语音识别（ASR）系统中校正音频中的分组丢失的系统和方法。 如本文所示，分组丢失校正发生在识别阶段，而不修改在训练期间产生的任何声学模型。 因此，在没有丢包的情况下，ASR引擎的行为不会改变。 为了实现这一点，实际输入信号可以被纠正，识别分数可以被归一化以考虑信号误差，并且可以使用使用来自先前帧和声学模型的信息的最佳估计方法来代替噪声信号。

公开(公告)号：US20170193988A1

公开(公告)日：2017-07-06

申请号：US14985560

申请日：2015-12-31

Applicant: Interactive Intelligence Group, Inc.

Inventor： Srinath Cheluvaraja ,  Ananth Nagaraja Iyer

IPC: G10L15/14 ,  G10L25/24 ,  G10L25/27 ,  G10L15/16

CPC classification number: G10L15/144 ,  G06N3/02 ,  G10L15/14 ,  G10L15/16 ,  G10L25/24 ,  G10L25/27

Abstract: A system and method are presented for neural network based feature extraction for acoustic model development. A neural network may be used to extract acoustic features from raw MFCCs or the spectrum, which are then used for training acoustic models for speech recognition systems. Feature extraction may be performed by optimizing a cost function used in linear discriminant analysis. General non-linear functions generated by the neural network are used for feature extraction. The transformation may be performed using a cost function from linear discriminant analysis methods which perform linear operations on the MFCCs and generate lower dimensional features for speech recognition. The extracted acoustic features may then be used for training acoustic models for speech recognition systems.