arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1906.08287cs.CL
收藏

深度时序模型的嵌入时间表达式

Embedding time expressions for deep temporal ordering models

The University of Texas at Austin(德克萨斯大学奥斯汀分校)

Tanya Goyal, Greg Durrett

摘要:

数据驱动模型在推断文本中事件的时间顺序方面表现出最先进的性能。然而,这些模型往往忽略了明确的时间信号,如日期和时间窗口。基于规则的方法可以用来识别这些时间表达式(时间轴)之间的时间联系,但它们无法捕捉时间轴与事件的相互作用,并且难以与神经网络模型的分布式表示集成。在本文中,我们引入了一个框架,通过学习预训练的模型来嵌入时间,从而将时间意识注入到这些模型中。我们生成由时间对组成的合成数据,然后训练一个字符LSTM学习嵌入和分类时间的时间关系。我们评估了这些嵌入在事件时间排序的强神经模型的背景下的实用性,并显示了在MATRES数据集上的性能略有提高,以及在具有更频繁的事件-时间交互的自动收集数据集上的更大收益。

英文摘要:

Data-driven models have demonstrated state-of-the-art performance in inferring the temporal ordering of events in text. However, these models often overlook explicit temporal signals, such as dates and time windows. Rule-based methods can be used to identify the temporal links between these time expressions (timexes), but they fail to capture timexes' interactions with events and are hard to integrate with the distributed representations of neural net models. In this paper, we introduce a framework to infuse temporal awareness into such models by learning a pre-trained model to embed timexes. We generate synthetic data consisting of pairs of timexes, then train a character LSTM to learn embeddings and classify the timexes' temporal relation. We evaluate the utility of these embeddings in the context of a strong neural model for event temporal ordering, and show a small increase in performance on the MATRES dataset and more substantial gains on an automatically collected dataset with more frequent event-timex interactions.

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。

↑