arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.06474cs.CLcs.AI
收藏

基于LSTM的跨话语信息表示的Transformer语言模型

Transformer Language Models with LSTM-based Cross-utterance Information Representation

Cambridge University Engineering Dept.(剑桥大学工程系)

G. Sun, C. Zhang, P. C. Woodland

摘要:

跨话语信息的有效结合有可能改进用于自动语音识别(ASR)的语言模型(LM)。为了为Transformer LM(TLM)提取更强大和鲁棒的跨话语表示,本文提出了在长短期记忆(LSTM)LM中使用隐藏状态的R-TLM。为了对交叉话语信息进行编码,R-TLM将LSTM模块与部分Transformer块中的分段递归结合在一起。除了LSTM模块输出之外,还研究了使用绕过LSTM模块的融合层的快捷连接。该系统在AMI会议语料库、Eval 2000和RT 03电话会话评估集上进行了评估。在AMI评估集Eval 2000和RT 03上,最佳R-TLM在单话语TLM基线上分别实现了0.9%、0.6%和0.8%的绝对WER降低,在强跨话语TLM基线上分别实现了0.5%、0.3%和0.2%的绝对WER降低。显著性检验进一步支持Eval 2000和RT 03的改进。R-TLM被发现在识别错误更容易发生的单词上有更好的LM分数。R-TLM WER可以通过利用LSTM-LM的插值来进一步减小。

英文摘要:

The effective incorporation of cross-utterance information has the potential to improve language models (LMs) for automatic speech recognition (ASR). To extract more powerful and robust cross-utterance representations for the Transformer LM (TLM), this paper proposes the R-TLM which uses hidden states in a long short-term memory (LSTM) LM. To encode the cross-utterance information, the R-TLM incorporates an LSTM module together with a segment-wise recurrence in some of the Transformer blocks. In addition to the LSTM module output, a shortcut connection using a fusion layer that bypasses the LSTM module is also investigated. The proposed system was evaluated on the AMI meeting corpus, the Eval2000 and the RT03 telephone conversation evaluation sets. The best R-TLM achieved 0.9%, 0.6%, and 0.8% absolute WER reductions over the single-utterance TLM baseline, and 0.5%, 0.3%, 0.2% absolute WER reductions over a strong cross-utterance TLM baseline on the AMI evaluation set, Eval2000 and RT03 respectively. Improvements on Eval2000 and RT03 were further supported by significance tests. R-TLMs were found to have better LM scores on words where recognition errors are more likely to occur. The R-TLM WER can be further reduced by interpolation with an LSTM-LM.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑