Weighted Argumentation for Web Science[go to overview]
Weighted Argumentation frameworks are a computationally efficient tool to represent and reason over argumentation problems.Natural application examples include decision problems and social media analysis. We will focus on web applications and discuss some current research on learning weighted argumentation frameworks from web content like online discussions and product reviews.
This allows, for example, learning about the public opinion about political topics and aggregating user opinions. A flawless fully automatic extraction of argumentation graphs from arbitrary domains is currently out of reach. However, recent work demonstrates how the state-of-the-art can already improve performance and explainability of conventional methods in domains like review aggregation, fake news detection and information filtering.
15.10.19 - 10:15