VIP内容

近年来,网络上的虚假信息越来越猖獗。研究界提出方法检测假信息(各种子任务)有用。然而,大多数提出的系统是基于深度学习技术,这些技术是针对特定领域进行微调的,很难解释和产生人类可理解的结果。这限制了它们的适用性和采纳,因为它们只能由经过挑选的专家观众在非常特定的设置中使用。在本文中,我们提出了一个基于可信度审查(CRs)的核心概念的架构,可用于构建协作进行虚假信息检测的分布式机器人网络。CRs作为构建模块来组成 (i) Web内容的图表,(ii) 现有的可信度信号——经过事实核查的声明和网站的声誉评论——和(iii) 自动计算的评论。我们在http URL的轻量级扩展之上实现这个架构,并提供用于语义相似度和立场检测的通用NLP任务。对现有社交媒体帖子、假新闻和政治演讲数据集的评估显示,与现有系统相比,这些数据集有几个优势: 可扩展性、领域独立性、可组合性、可解释性和来源透明性。此外,我们不需要微调就能获得比赛结果,并在Clef'18上建立了一种新的艺术状态。

https://www.zhuanzhi.ai/paper/041cde1a69557e5a8b320366c3179253

成为VIP会员查看完整内容
0
12

最新内容

This paper summarizes our participation in the SMART Task of the ISWC 2020 Challenge. A particular question we are interested in answering is how well neural methods, and specifically transformer models, such as BERT, perform on the answer type prediction task compared to traditional approaches. Our main finding is that coarse-grained answer types can be identified effectively with standard text classification methods, with over 95% accuracy, and BERT can bring only marginal improvements. For fine-grained type detection, on the other hand, BERT clearly outperforms previous retrieval-based approaches.

0
0
下载
预览

最新论文

This paper summarizes our participation in the SMART Task of the ISWC 2020 Challenge. A particular question we are interested in answering is how well neural methods, and specifically transformer models, such as BERT, perform on the answer type prediction task compared to traditional approaches. Our main finding is that coarse-grained answer types can be identified effectively with standard text classification methods, with over 95% accuracy, and BERT can bring only marginal improvements. For fine-grained type detection, on the other hand, BERT clearly outperforms previous retrieval-based approaches.

0
0
下载
预览
Top