Computational Linguistics Colloquium

Thursday, 25 July 2013, 16:15
Conference Room, Building C7.4

Machine Learning of Personality Assignment and Deception Detection

Computational Stylometry is an area in text analytics in which information about the author of a text is extracted rather than about the contents of the text. Examples include gender, age, and personality detection, and in the limit assignment of the author's identity (authorship attribution). Parallel to this research, there is an increasing interest in deception detection; deciding on a text (e.g. a product review) whether it is true or deceptive. Interestingly, similar text features are correlated with both personality and deception. In this talk, I will explain the state of the art of both detection tasks in a machine learning context, and explain how machine learning methods such as meta-learning (a variant of ensemble methods) and joint learning can provide a new perspective on these tasks and on unraveling their interaction.

Walter Daelemans is a professor of computational linguistics at the University of Antwerp where he heads the CLiPS Computational Linguistics group. His current research interests are in machine learning of language, deep text understanding, and computational stylometry.

If you would like to meet with the speaker, please contact Dietrich Klakow.