BioNLP dataset

About Complex mentions:

• The following lines from a review paper Recognizing Complex Entity Mentions: A Review and Future Directions
• Three types of complex mentions: nested, overlapping and discontinuous
• GENIA (Kim et al., 2003) only contains nested entity mention.
• CADEC (Karimi et al., 2015) and SemEval2014 (Pradhan et al., 2014) contain
  overlapping and discontinuous mentions.
• ACE 1 and NNE (Ringland, 2016) are newswire corpora with nested entity mentions.

SemEval2014:

• There are two sub tasks in this task. The first one is to extract the overlapping and discontinuous mentions; The second is to postprocess the extracted entities, like norminalizing some tokens, add some missing tokens, etc.
• [Please refer the cite for detail, and the original paper is here

CADEC

• This data contains overlapping and discontinuous mentions and adverse drug events. Please refer the paper for details. CADEC: A corpus of adverse drug event annotations
• Here is the original URL for download the CADEC dataset download URL

BioNLP-Share Task 2011

• There are several tasks including main task and sub task.
  • GENIA, nested mentions and biomedical event
• Here is the web site

References

• Joint Extraction of Entities and Relations Based on a Novel Graph Scheme
• Joint Extraction of Entities and Overlapping
  Relations Using Position-Attentive Sequence Labeling
• Learning to Recognize Discontiguous Entities
• An Effective Transition-based Model for Discontinuous NER