Becoming smarter about credible news
summary
Thomson Reuters shows a live system that clusters real-time tweets and articles into user-defined channels, extracts topics and concepts, scores each cluster for credibility with a veracity score, and exposes indexed, searchable feeds and analytics for journalists and financial clients. The system reports 84% accuracy separating fake versus real items for clusters with at least five tweets and 78% accuracy distinguishing rumors from factual assertions. The pipeline uses big data plus machine learning for lead discovery, source verification, and automated filtering that can feed trading desks or newsroom monitors.
AI-written summary · about
related talks
34mApplying Machine Learning Techniques to Detect and Understand Fake News- 34mApplying Machine Learning Techniques to Detect and Understand Fake News
03mServing Billions of Personalized News Feeds with AI
10mReal-Time Fraud Prevention at Scale
21m"What's in the news?" #bbuzz
11mIdentifying Disinformation Websites Using Infrastructure Features
42mScaling to Billions of Real-time Recommendations Daily
30mProcessing Threat Reports at Scale Using AI and ML: Expectations and Reality