MSc Thesis Defence: Raluca Alexandra Fetic
Efficient neural encoding of large texts for stance detection
We report an analysis and a set of experiments build around investigating how to efficiently encode large texts for stance detection. By modelling this as a fact checking task, we attempt to automatically asses the veracity of claims based on evidence documents related to them. We examine theory in natural language understanding and deep learning literature relevant to this problem and discuss various algorithms.
Speaker: Raluca Alexandra Fetic
Censor: Zeljko Agic
Supervisor: Isabelle Augenstein