AAAI 2022 finer grained false information detection

False information in social media （misinformation） It refers to the wrong information spread through social networking platforms 、 Inaccurate information . At this stage, the detection of false information can be seen from 3 Implement in three aspects ：（1） Fake news detection （fake news detection）, That is to judge the authenticity of news content with long text , Usually assisted by relevant comments ;（2） Rumor detection （rumor detection）, It is mostly the communication analysis of short texts on social platforms , That is, through people's comments and forwarding of information on social networks , Analyze the dissemination characteristics of false information and the opinions of relevant discussions ;（3） Proof of fact （fact verification）, That is to retrieve the content involved in the information through the search engine , With the supplement of factual information, judge whether the original judgment is true or not . This article will introduce 3 In AAAI 2022 Related work in false information detection in .

• False news detection based on fine-grained reasoning

Towards Fine-Grained Reasoning for Fake News Detection

Address of thesis : 

https://www.aaai.org/AAAI22Papers/AAAI-12332.JinY.pdf

This article aims at the task of false news detection , By better reflecting the logical reasoning process of human thinking and realizing the modeling of subtle clues , Promote false news detection towards fine-grained reasoning .

• Dual dynamic graph neural network for rumor detection

DDGCN: Dual Dynamic Graph Convolutional Networks for Rumor Detection on Social Media 

Address of thesis :

 https://www.aaai.org/AAAI22Papers/AAAI-6370.SunM.pdf

This article aims at the task of rumor detection , According to the time stage of forwarding the comments contained in the rumor information , Two dynamic graphs, message propagation network and knowledge entity evolution network, are constructed , Integrate timing and propagation structure , Make the overall representation more logical .

• Interpretable fact verification reasoning paradigm

LOREN: Logic-Regularized Reasoning for Interpretable Fact Verification

Address of thesis : 

https://www.aaai.org/AAAI22Papers/AAAI-8344.ChenJ.pdf

This article aims at the task of fact verification , Break down the validation of the entire statement into phrase level validation , Verify the results by integrating the fine-grained information in the statement , More intuitive display of model reasoning process , It can also lock in factual errors faster .