Unsupervised Learning of Acoustic Sub-word Units
Unsupervised Learning of Acoustic Sub-word Units
复制标题
声学子词单元的无监督学习
DOI：
--
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
Emmanuel Dupoux
中科院分区：
文献类型：
--
作者：
Balakrishnan Varadarajan;S. Khudanpur;Emmanuel Dupoux
精确的无监督学习的音素的语言直接从语音演示通过一个算法的联合无监督学习的拓扑结构和参数的隐马尔可夫模型（HMM）;状态和短状态序列，通过这个HMM对应于学习的子字单元。该算法，最初提出了一个给定的音素集内的allophonic变化的无监督学习，已被改编为学习没有任何知识的音素。还提出了一种评估方法，从而对齐到测试话语的状态序列以自动的方式转换到音素序列，并与其手动转录进行比较。超过85%的音素识别准确率被证明为从流利的，大词汇量的语音中进行依赖于说话者的学习。
Accurate unsupervised learning of phonemes of a language directly from speech is demonstrated via an algorithm for joint unsupervised learning of the topology and parameters of a hidden Markov model (HMM); states and short state-sequences through this HMM correspond to the learnt sub-word units. The algorithm, originally proposed for unsupervised learning of allophonic variations within a given phoneme set, has been adapted to learn without any knowledge of the phonemes. An evaluation methodology is also proposed, whereby the state-sequence that aligns to a test utterance is transduced in an automatic manner to a phoneme-sequence and compared to its manual transcription. Over 85% phoneme recognition accuracy is demonstrated for speaker-dependent learning from fluent, large-vocabulary speech.