In this presentation, I will provide an overview of the first stages of my PhD, 
starting with my initial ideas and how the developed to my current research plans. 

My PhD research is directed towards improving stochastic disambiguation models for 
deep grammars.
  Currently, we are running experiments directed towards finding the "state of the art" 
of disambiguation models for the ERG. Future plans are to extend these experiments to 
open text, and improve on disambiguation models, particularly when shallow tools are used. 
One of the main challenges for these goals is the absence of (parser independent) training 
and test data. Improving on existing models, as well as getting a reliable indication of 
performance on open text is impossible without such data. This is therefore the next problem 
that will be addressed in my research. I will look at syntactic annotations and outputs of 
different parsers. The goal is to map different outputs to a common representation so that 
treebanks developed for different parsers can be used as training and test data for the ERG. 
Naturally, this approach would also allow to compare performance of different parsers.