PaperWeekly
文章平均质量分 87
静待花开s0
If you find a path with no obstacles, it probably doesn't lead anywhere.
展开
-
论文阅读笔记《CAsT-19: A Dataset for Conversational Information Seeking》
CAsT-19: A Dataset for Conversational Information SeekingSIGIR-2020paper link:https://dl.acm.org/doi/abs/10.1145/3397271.3401206Conversational search benchmark at TREC:http://www.treccast.ai/1.简介CAsT-19是一个支持对话信息检索(对话式检索)研究的新数据集,该数据集共有来自TREC Complex An原创 2021-01-28 22:32:16 · 563 阅读 · 0 评论 -
机器阅读理解(Neural Machine Reading Comprehension)综述,相关方法及未来趋势
Neural Machine Reading Comprehension:Methods and TrendsAuthor:Shanshan Liu , Xin Zhang , Sheng Zhang , Hui Wang , Weiming ZhangPDF: https://arxiv.org/abs/1907.011180.写在前面:机器阅读理解(MRC)需要机器回答基于给定上下文的问题,在过去几年里,随着各种深度学习技术的融合,它越来越受到关注。本文发表于2019年,主要从以下几个方面对该领原创 2021-01-21 22:36:02 · 1247 阅读 · 0 评论 -
论文阅读笔记——CDL: Curriculum Dual Learning for Emotion-Controllable Response Generation
CDL: Curriculum Dual Learning for Emotion-Controllable Response GenerationAuthor:Lei Shen Yang Feng会议:ACL2020paper:https://arxiv.org/abs/2005.00329简介:我们知道,在开放域对话的回复生成过程中,添加情感更能提高对话质量,而目前现有的方法主要是通过在标准交叉熵损失中加入正则化项增加情感表达,从而影响训练过程,但缺乏对内容一致性的进一步考虑,同时加剧了safe原创 2021-01-14 22:29:00 · 697 阅读 · 0 评论