arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2203.15917cs.CL
收藏

加权状态转换器与语言模型的浅层融合文本规范化

Shallow Fusion of Weighted Finite-State Transducer and Language Model for Text Normalization

NVIDIA(英伟达)

Evelina Bakhturina, Yang Zhang, Boris Ginsburg

摘要:

生产中的文本规范化(TN)系统在很大程度上是基于规则的,使用加权有限状态转换器(WFST)。然而,当规范化形式是上下文相关的时,基于WFST的系统与模糊输入作斗争。另一方面,神经文本规范化系统可以考虑上下文,但他们遭受不可恢复的错误,并需要标记的规范化数据集,这是很难收集。我们提出了一种新的混合方法,结合了基于规则和神经系统的好处。首先,非确定性WFST输出所有标准化候选,然后神经语言模型选择最佳候选-类似于自动语音识别的浅层融合。虽然WFST防止不可恢复的错误,但语言模型解决了上下文歧义。该方法很容易扩展,我们表明它是有效的。它实现了可比或更好的结果比现有的国家的最先进的TN模型。

英文摘要:

Text normalization (TN) systems in production are largely rule-based using weighted finite-state transducers (WFST). However, WFST-based systems struggle with ambiguous input when the normalized form is context-dependent. On the other hand, neural text normalization systems can take context into account but they suffer from unrecoverable errors and require labeled normalization datasets, which are hard to collect. We propose a new hybrid approach that combines the benefits of rule-based and neural systems. First, a non-deterministic WFST outputs all normalization candidates, and then a neural language model picks the best one -- similar to shallow fusion for automatic speech recognition. While the WFST prevents unrecoverable errors, the language model resolves contextual ambiguity. The approach is easy to extend and we show it is effective. It achieves comparable or better results than existing state-of-the-art TN models.

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。

↑