arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.01239cs.CL
收藏

从Wikipedia类别中挖掘知识用于自然语言推理

Mining Knowledge for Natural Language Inference from Wikipedia Categories

Toyota Technological Institute at Chicago(芝加哥丰田理工学院);Rutgers University(罗格斯大学);University of Chicago(芝加哥大学)

Mingda Chen, Zewei Chu, Karl Stratos, Kevin Gimpel

摘要:

精确的词汇蕴涵(LE)和自然语言推理(NLI)通常需要大量昂贵的注释。为了减轻对标记数据的需求,我们引入了WikiNLI:一个用于提高NLI和LE任务模型性能的资源。它包含428,899对短语,这些短语是从维基百科中自然注释的类别层次结构中构建的。我们表明,我们可以通过在WikiNLI上对BERT和RoBERTa进行预训练并将模型转移到下游任务上来改进强基线。我们与从其他知识库(如WordNet和Wikidata)中提取的短语进行了系统的比较,发现在WikiNLI上进行预训练可以获得最佳性能。此外,我们还用其他语言构建了WikiNLI,并表明对它们进行预训练可以提高相应语言的NLI任务的性能。

英文摘要:

Accurate lexical entailment (LE) and natural language inference (NLI) often require large quantities of costly annotations. To alleviate the need for labeled data, we introduce WikiNLI: a resource for improving model performance on NLI and LE tasks. It contains 428,899 pairs of phrases constructed from naturally annotated category hierarchies in Wikipedia. We show that we can improve strong baselines such as BERT and RoBERTa by pretraining them on WikiNLI and transferring the models on downstream tasks. We conduct systematic comparisons with phrases extracted from other knowledge bases such as WordNet and Wikidata to find that pretraining on WikiNLI gives the best performance. In addition, we construct WikiNLI in other languages, and show that pretraining on them improves performance on NLI tasks of corresponding languages.

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。

↑