RDoC Task 2019
Information Retrieval and Sentence Extraction on Mental Health using Research Domain Criteria
RDOC Task 2019 is closed now. However, the datasets are available here.
RDoC Task at BioNLP-OST 2019
The RDoC Task is part of this year's BioNLP Open Shared Tasks (BioNLP-OST: http://2019.bionlp-ost.org) .
There are two subtasks. Task 1 is information retrieval task in which the participants are required to develop models for retrieving PubMed abstracts relevant to several mental health related categories. Task 2 is a sentence extraction task in which participants are required to extract the most relevant sentence for each category from the given relevant abstract.
BioNLP-OST workshop will be collocated with EMNLP-IJCNPL in Hong-Kong later this year. And the evaluation of the RDoC task will be presented at this workshop.
Timeline
- March 18: Task Registration open
- March 25: Sample data release
- April 15: Training data (batch 1) release
- April 22: Training data (batch 2) release
- June 12: Test data (batch 1) release
- June 13: Webserver for submission available
- June 15: Test data (batch 2) release
- June 19: Evaluation complete - See results here.
- Aug. 19: Workshop paper submission
- Nov. 3 or 4: BioNLP-OST workshop
Introduction and Motivation
Research Domain Criteria (RDoC) is a recently developed comprehensive framework for describing mental disorders introduced by National Institute of Mental Health (NIMH). It provides a novel mechanism for describing mental illnesses using multiple dimensions in contrast to traditional one-dimensional symptom-based approaches .
To reach the full potential of RDoC, mental health researchers and clinicians should be able to quickly retrieve and extract information from biomedical articles related to RDoC categories (called "constructs"). However, there does not exist a publicly available biomedical literature annotated with RDoC constructs. One of the main reasons for the above is the tedious and time-consuming nature of the human-based manual curation of biomedical literature with RDoC annotations.
As a solution to the above problem, we invite text mining teams to develop models for information retrieval and sentence extraction with RDoC.
RDoC Task Objectives
RDoC task is a combination of two subtasks on information retrieval and sentence extraction.
- Task 1: retrieving PubMed Abstracts related to RDoC constructs
- Task 2: Extracting the most relevant sentences for RDoC construct from a relevant abstract.
More details are given in Task page.
Organizers
- Indika Kahanda, Montana State University, indika.kahanda@montana.edu
- Mohammad Anani, Montana State University, mohamed.anani@live.com
- Nazmul Kazi, Montana State University, kazinazmul.hasan@montana.edu
- Matt Kuntz, MSU and National Alliance of Mental Illness Montana, matthew.kuntz@montana.edu