Lattice segmentation and minimum Bayes risk discriminative training
Lattice segmentation and minimum Bayes risk discriminative training
复制标题
DOI：
10.21437/eurospeech.2003-573
复制
发表时间：
2003-09
期刊：
影响因子：
--
通讯作者：
Vlasios Doumpiotis;Stavros Tsakalidis;W. Byrne
中科院分区：
文献类型：
--
作者：
Vlasios Doumpiotis;Stavros Tsakalidis;W. Byrne
建模方法，将歧视性的训练程序，分段最小贝叶斯风险解码（SMBR）。SMBR用于将一般自动语音识别（ASR）系统产生的格分割成涉及小的易混淆单词集合的单独决策问题的序列。我们讨论了两种方法，将这些分割的格子在歧视性训练。我们研究了使用声学模型专门区分这些类中的竞争词，然后在随后的SMBR重新评分通过。细化的搜索空间，允许使用专门的判别模型被证明是一个改进与传统训练的判别模型rescoring。
Modeling approaches are presented that incorporate discriminative training procedures in segmental Minimum Bayes-Risk decoding (SMBR). SMBR is used to segment lattices produced by a general automatic speech recognition (ASR) system into sequences of separate decision problems involving small sets of confusable words. We discuss two approaches to incorporating these segmented lattices in discriminative training. We investigate the use of acoustic models specialized to discriminate between the competing words in these classes which are then applied in subsequent SMBR rescoring passes. Refinement of the search space that allows the use of specialized discriminative models is shown to be an improvement over rescoring with conventionally trained discriminative models.