arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1905.07129cs.CL
收藏

ERNIE:用信息实体增强的语言表示

ERNIE: Enhanced Language Representation with Informative Entities

Tsinghua University(清华大学);Huawei Noah’s Ark Lab(华为诺亚方舟实验室)

Zhengyan Zhang, Xu Han, Zhiyuan Liu, Xin Jiang, Maosong Sun, Qun Liu

摘要:

在大规模语料库上预训练的神经语言表示模型(如BERT)可以很好地从纯文本中捕获丰富的语义模式,并进行微调,以持续提高各种NLP任务的性能。然而,现有的预训练语言模型很少考虑纳入知识图谱(KG),这可以提供丰富的结构化知识事实,以更好的语言理解。我们认为,知识库中的信息实体可以增强语言表征与外部知识。在本文中,我们利用大规模的文本语料库和知识库训练一个增强的语言表示模型(ERNIE),它可以充分利用词汇,句法和知识信息的同时。实验结果表明,ERNIE在各种知识驱动的任务上取得了显著的改进,同时在其他常见的NLP任务上与最先进的模型BERT相当。本文的源代码可以从https://github.com/thunlp/ERNIE获得。

英文摘要:

Neural language representation models such as BERT pre-trained on large-scale corpora can well capture rich semantic patterns from plain text, and be fine-tuned to consistently improve the performance of various NLP tasks. However, the existing pre-trained language models rarely consider incorporating knowledge graphs (KGs), which can provide rich structured knowledge facts for better language understanding. We argue that informative entities in KGs can enhance language representation with external knowledge. In this paper, we utilize both large-scale textual corpora and KGs to train an enhanced language representation model (ERNIE), which can take full advantage of lexical, syntactic, and knowledge information simultaneously. The experimental results have demonstrated that ERNIE achieves significant improvements on various knowledge-driven tasks, and meanwhile is comparable with the state-of-the-art model BERT on other common NLP tasks. The source code of this paper can be obtained from https://github.com/thunlp/ERNIE.

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。

↑