An efficient two-pass search algorithm using word trellis index

An efficient two-pass search algorithm using word trellis index
复制标题

一种利用词网格索引的高效两遍搜索算法

DOI:
10.21437/icslp.1998-626
复制
发表时间:
1998
期刊:
5th International Conference on Spoken Language Processing (ICSLP 1998)
影响因子:
--
通讯作者:
S. Doshita
S. Doshita
中科院分区:
--
文献类型:
--
作者:
Akinobu Lee;Tatsuya Kawahara;S. Doshita

文献摘要

参考文献

被引文献

相似文献

我们提出了一个有效的两遍搜索算法LVCSR。代替传统的词图,第一个初步通过生成“词网格索引”,在每个时间帧内跟踪波束内所有幸存的词假设。由于它非确定性地表示所有找到的词边界,因此我们可以(1)在第二次搜索时获得准确的依赖于词对的假设,以及(2)避免在第一次搜索时进行昂贵的词对近似。第二遍执行高效堆栈解码搜索,其中索引被称为预测单词列表和索引。在5,000字的日语听写任务上的实验结果表明,与词图方法相比,该方法在保持较高准确率的同时,内存开销小于1/10。最后,通过处理词间上下文相关性,我们实现了5.6%的词错误率。
We propose an e cient two-pass search algorithm for LVCSR. Instead of conventional word graph, the rst preliminary pass generates \word trellis index", keeping track of all survived word hypotheses within the beam every time-frame. As it represents all found word boundaries non-deterministically, we can (1) obtain accurate sentence-dependent hypotheses on the second search, and (2) avoid expensive word-pair approximation on the rst pass. The second pass performs an e cient stack decoding search, where the index is referred to as predicted word list and heuristics. Experimental results on 5,000-word Japanese dictation task show that, compared with the word-graph method, this trellis-based method runs with less than 1/10 memory cost while keeping high accuracy. Finally, by handling inter-word context dependency, we achieved the word error rate of 5.6%.
T.Kawahara,T.Kobayashi,K.Takeda,N.Minematsu,K.Itou,M.Yamamoto,A.Yamada,T.Utsuro,K.Shikano:“用于日语大词汇连续语音识别的可共享软件存储库”Proc。
DOI: --
发表时间: --
期刊:
影响因子: --
作者:
通讯作者: --