Language modeling and transcription of the TED corpus lectures
Language modeling and transcription of the TED corpus lectures
复制标题
TED 语料库讲座的语言建模和转录
DOI：
10.1109/icassp.2003.1198760
复制
发表时间：
2003
期刊：
影响因子：
--
通讯作者：
M. Cettolo
中科院分区：
文献类型：
--
作者：
Erwin Leeuwis;Marcello Federico;M. Cettolo
无论是在声学方面还是在语言建模方面，转录讲座都是一项具有挑战性的任务。在这项工作中，我们介绍了我们在TED语料库中自动转录演讲的第一个结果，该语料库最近由ELRA和LDC发布。特别是，我们将精力集中在语言建模上。基线声学和语言模型分别使用8小时的TED文本和不同类型的文本：会议记录、演讲记录和对话演讲记录。然后，通过利用不同类型的信息来研究语言模型对单个说话人的适应性：演讲的自动抄本，演讲的标题，摘要，最后是论文。在最后一种情况下，获得了39.2%的WER。
Transcribing lectures is a challenging task, both in acoustic and in language modeling. In this work, we present our first results on the automatic transcription of lectures from the TED corpus, recently released by ELRA and LDC. In particular, we concentrated our effort on language modeling. Baseline acoustic and language models were developed using respectively 8 hours of TED transcripts and various types of texts: conference proceedings, lecture transcripts, and conversational speech transcripts. Then, adaptation of the language model to single speakers was investigated by exploiting different kinds of information: automatic transcripts of the talk, the title of the talk, the abstract and, finally, the paper. In the last case, a 39.2% WER was achieved.