Asynchronous-Transition Hidden Markov Model with State-Tying across Time for Automatic Speech Recognition

用于自动语音识别的具有跨时间状态绑定的异步转移隐马尔可夫模型

• 批准号: 12680375
• 负责人: SHIMODAIRA Hiroshi
• 金额: $ 2.18万
• 依托单位: Japan Advanced Institute of Science and Technology
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2000
• 资助国家: 日本
• 起止时间: 2000 至 2002
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-12680375/
• 关键词: hidden Markov model; HMM; asynchronous-transition; AT-HMM; 非同期遷移型HMM; 時間方向共有; 特定話者音声認識; 複数軌道モデル; 特徴量別音素環境依存モデル; 特徴量依存音素環境クラスタリング

This project aimed to improve acoustic models for speech recognition systems. The state-of-the-art hidden Markov model (HMM) based acoustic models usually treat the acoustic features as a chain of stationary signal sources. The observed values of these features are represented by vectors. We assumed that they might be better modeled by individual vector components. We discussed two methods based on this assumptionIn the first method, wearied to model asynchronous changes of individual acoustic vector components. Conventional HMM implicitly assumes that individual components change their statistical properties simultaneously. This assumption might be not true. Temporally changing patterns of individual acoustic components do not necessarily synchronize with beach other. We proposed a new HMM that allowed asynchronous state transitions between individual vector components. We demonstrated that this new HMM outperformed the conventional HMM in speaker-dependent speech recognition taskIn the second method, we tried to model phoneme context dependency of individual acoustic vector components. Conventional parameter tying techniques provide a common tying structure for all vector components, no matter how different is their individual components complexity and phoneme context dependency. In this discussion, we proposed a new parameter tying technique that allowed to have distinct tying structures for each component. Our experimental results showed that proposed HMM with feature-depended tying worked better than conventional HMM with a common tying

该项目旨在改进语音识别系统的声学模型。基于隐马尔可夫模型（HMM）的声学模型通常将声学特征视为一系列平稳信号源。这些特征的观测值由向量表示。我们假设它们可能更好地由单个矢量分量建模。在此基础上讨论了两种方法，第一种方法是对单个声矢量分量的异步变化进行建模。传统的HMM隐含地假设各个分量同时改变它们的统计特性。这个假设可能不正确。各个声学分量的时间变化模式不一定与其他海滩同步。我们提出了一种新的HMM，允许异步状态之间的转换个别向量分量。我们证明，这种新的HMM优于传统的HMM在说话人相关的语音识别taskIn第二种方法，我们试图模拟音素上下文依赖的个别声学矢量分量。传统的参数绑定技术为所有矢量分量提供了一个公共的绑定结构，无论它们的各个分量的复杂性和音素上下文依赖性有多么不同。在这次讨论中，我们提出了一个新的参数绑定技术，允许有不同的绑定结构的每个组件。我们的实验结果表明，提出的HMM与特征依赖的搭售工作优于传统的HMM与共同的搭售

项目成果

S.Matsuda, M.Nakai, H.Shimodaira, S.Sagayama: "Generation of phoneme environment clusters with multiple trajectories"The 2001, Autumn Meeting of The Acoustical Society of Japan, (in Japanese). 1-1-10 (Oct). 19-20 (2001)

S.Matsuda、M.Nakai、H.Shimodaira、S.Sagayama：“具有多个轨迹的音素环境簇的生成”，日本声学学会 2001 年秋季会议（日语）。

松田 繁樹: "音素環境クラスタリングを基礎としたマルチパス音響モデルの自動生成"日本音響学会2002年秋季研究発表会講演論文集. 81-82 (2002)

Shigeki Matsuda：“基于音素环境聚类的多路径声学模型的自动生成”日本声学学会 2002 年秋季研究会议论文集 81-82 (2002)。

S.Matsuda, M.Nakai, H.Shimodaira, S.Sagayama: "Feature-dependent Allophone Clustering"Proc.International Conference of Spoken Language Processing (CSLP2000). (Oct). 413-416 (2000)

S.Matsuda、M.Nakai、H.Shimodaira、S.Sagayama：“特征相关的音位变体聚类”Proc.国际口语处理会议（CSLP2000）。

S.Matsuda: "Feature-dependent Allophone Clustering"Proc. International Conference on Spoken Language Processing (IC-SLP2000). 1. 413-416 (2000)

S.Matsuda：“特征相关的变体聚类”Proc。

松田 繁樹: "複数の特徴ベクトル軌道を持つ環境依存音素クラスタの生成"日本音響学会2001年秋季研究発表会講演論文集. 19-20 (2001)

Shigeki Matsuda：“具有多个特征向量轨迹的环境相关音素簇的生成”日本声学学会 2001 年秋季研究会议记录 19-20 (2001)。