WSDMS: Debunk Fake News via Weakly Supervised Detection of Misinforming Sentences with Contextualized Social Wisdom

Keyphrases

• Fake News 100%
• Weakly Supervised Object Detection 100%
• Sentence-level 80%
• Misinformation 60%
• Social Media 40%
• Article-level 40%
• Training Data 20%
• News Articles 20%
• Existing State 20%
• Viral 20%
• Learning Approaches 20%
• Social Media Users 20%
• Multiple Instance Learning 20%
• Real-world Benchmark 20%
• Stance Detection 20%
• Falsehood 20%
• Rumor Verification 20%
• Rumor Stance 20%
• Stance Expression 20%

Computer Science

• Misinformation 100%
• Annotation 33%
• Multiple Instance 33%
• Learning Approach 33%
• Social Medium User 33%
• Instance Learning 33%
• Training Dataset 33%
• Stance Detection 33%