SiGHTSee - Simultaneous Generation
in Heidelberg: Text and Scene images
SiGHTSee at KES2011
Our recent paper submission has been accepted for the 15th Annual KES Conference in Kaiserslautern, Germany, September 12-14. The full paper, describing the multimodal generation architecture in our project, can be found here.
On March 22, Katarina Boland and Michael Roth will present the SiGHTSee project in a FRONTIER Symposium organised by Heidelberg University. The complete program can be found here.
Two project-related theses have been recently finished:
- Katarina Boland (2010): Entwicklung einer integrierten Repräsentation für synchrone sprachliche und visuelle Generierung. Magister's thesis, Heidelberg University. Supervisors: Anette Frank and Susanne Krömker.
- Maria Semenchuk (2011): Annotation und Analyse anaphorischer Bindung impliziter semantischer Rollen in Weginstruktionen. Bachelor's thesis, Heidelberg University. Supervisor: Anette Frank.
Participation in GIVE-2
We participated in the Second NLG Challenge on Generating Instructions in Virtual Environments (GIVE-2). The results were presented at the 6th International Natural Language Generation Conference (INLG 2010). See here for details.
SiGHTSee at COLING 2010
Michael Roth presented his work on aligning route representations and route directions at the 23rd International Conference on Computational Linguistics (Coling 2010) in Beijing, China. Check it out here.