We are working on modeling biological processes as structures, and using these structures for question answering in a reading comprehension task. We have shown that using process structures improves reading comprehension compared to shallower approaches.
The ProcessBank dataset used in the paper
Modeling Biological Processes for Reading Comprehension by Jonathan Berant, Vivek Srikumar, Pei-Chun Chen, Abby Vander Linden, Brittany Harding, Brad Huang, Peter Clark and Christopher D. Manning
is hosted by the Allen Institue for Artificial Intelligence. You can read about ProcessBank and download
the dataset
here.
You can download the code of our system here.
This is the code and data used in our EMNLP 2013 paper:
Learning Biological Processes with Global Constraint by Aju Thalappillil Scaria, Jonathan Berant, Mengqiu Wang, Peter Clark, Justin Lewis, Brittany Harding and Christopher D. Manning.
The dataset contains full annotation of 148 process descriptions split into a training set and a test set. Annotation was performed using the brat annotation tool and so it is easiest to look at the annotation by importing the dataset files to brat.
The software is open source, distributed under the BSD 3-clause license. You can download: