DDIExtraction 2013: Extraction of Drug-Drug Interactions from biomedical texts



Following the success of the DDIExtraction 2011 task, we invite your participation in the second DDIExtraction task, which will held at Semeval 2013, co-located with SEM-2013 and NAACL-2013 conferences in Atlanta, GA.

The DDIExtraction 2013 task concerns the recognition of drugs and extraction of drug-drug interactions that appear in biomedical literature. The definition of drug-drug interaction (DDI) is broadly described as a change in the effects of one drug by the presence of another drug (e.g. Acetazolamide reduces urinary excretion of quinidine and may enhance its effect.). The detection of DDIs is an important research area in patient safety since these interactions can become very dangerous and increase health care costs.

The goal of DDIExtraction 2013 is to provide a common framework for evaluation of information extraction techniques applied to the recognition of pharmacological substances and detection of drug-drug interactions from biomedical texts.  The DDIExtraction 2011 task concentrated efforts on the novel aspects of the DDI extraction task, the drug recognition was assumed and the annotations for drugs were provided to the participants. This new DDIExtraction 2013 task will pursue the extraction of drug interactions as well as the recognition of drug names. Additionally, while the DDIExtraction 2011 shared task relied on the DDI corpus composed by texts describing DDIs from the DrugBank database, the DDIExtraction2013 also integrates MedLine abstracts in order to deal with different type of texts and language styles.

DDIExtraction2013 Shared Task is designed to address the extraction of DDIs as a whole, but divided into two subtasks to allow separate evaluation of the performance for different aspects of the problem. This shared task includes two challenges:

TASK 9.1: Recognition and classification of drug names.

TASK 9.2: Extraction of drug-drug interactions. The extraction of drug-drug interactions is a specific relation extraction task in biomedical literature. This task could be very appealing to groups studying PPI (protein-protein interaction) extraction because they could adapt their systems to extract drug-drug interactions.


Training dataset can be found from: http://www.cs.york.ac.uk/semeval-2013/task9/index.php?id=data


Important dates

Nov 15: Training dataset available for participants
Feb 15: Registration Deadline [for Task Participants]
Mar 1: Start of evaluation period
Mar 15: End of evaluation period (23:59, UTC-11)
Apr 9: Paper submission deadline
May 4: Camera ready


  • Isabel Segura-Bedmar, University Carlos III of Madrid, Spain
  • Paloma Martínez, University Carlos III of Madrid, Spain
  • María Herrero, University Carlos III of Madrid, Spain



If interested in the task please join the mailing list for updates.


RocketTheme Joomla Templates