arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2202.12160cs.CL
收藏

不完全话语重写的自我注意

Self-Attention for Incomplete Utterance Rewriting

Ping An Technology (Shenzhen) Co., Ltd.(平安科技(深圳)有限公司)

Yong Zhang, Zhitao Li, Jianzong Wang, Ning Cheng, Jing Xiao

摘要:

不完整话语重写(IUR)是近年来自然语言处理领域的一项重要任务,旨在为不完整的话语补充足够的语境信息以促进理解。本文提出了一种新的方法,即直接从Transformer的自注意权矩阵中提取共指和省略关系,而不是从词嵌入中提取共指和省略关系,并对原文进行相应的编辑,从而生成完整的话语。利用自注意力权重矩阵中的丰富信息,该方法在公开的IUR数据集上取得了有竞争力的结果。

英文摘要:

Incomplete utterance rewriting (IUR) has recently become an essential task in NLP, aiming to complement the incomplete utterance with sufficient context information for comprehension. In this paper, we propose a novel method by directly extracting the coreference and omission relationship from the self-attention weight matrix of the transformer instead of word embeddings and edit the original text accordingly to generate the complete utterance. Benefiting from the rich information in the self-attention weight matrix, our method achieved competitive results on public IUR datasets.

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。

↑