Clinical TempEval 2016

Event Notification Type: 
Call for Papers
Sunday, 10 January 2016 to Sunday, 31 January 2016
Contact: 
Steven Bethard
Leon Derczynski
Guergana Savova
James Pustejovsky
Marc Verhagen

Clinical TempEval 2016 follows in the footsteps of i2b2 2012 and Clinical TempEval 2015 in bringing shared tasks on timeline extraction to the clinical domain. Participating systems receive as input text such as:

"April 23, 2014: The patient did not have any postoperative bleeding so we will resume chemotherapy with a larger bolus on Friday even if there is slight nausea."

And produce as output the times, events and temporal relations extracted from the text:

Times:
* April 23, 2014: type=DATE, value=2014-04-23
* postoperative: type=PREPOSTEXP
* Friday: type=DATE, value=2014-04-25

Events:
* bleeding: type=N/A, degree=N/A, polarity=NEG, modality=ACTUAL
* resume: type=ASPECTUAL, degree=N/A, polarity=POS, modality=ACTUAL
* chemotherapy: type=N/A, degree=N/A, polarity=POS, modality=ACTUAL
* bolus: type=N/A, degree=N/A, polarity=POS, modality=ACTUAL
* nausea: type=N/A, degree=LITTLE, polarity=POS, modality=HYPOTHETICAL

Relations:
* bleeding BEFORE DocumentCreationTime
* resume AFTER DocumentCreationTime
* chemotherapy AFTER DocumentCreationTime
* bolus AFTER DocumentCreationTime
* nausea AFTER DocumentCreationTime
* postoperative CONTAINS bleeding
* Friday CONTAINS resume
* Friday CONTAINS bolus

== What's New ==

Clinical TempEval 2016 will provide a new test set for the evaluation, so the training and test sets from Clinical TempEval 2015 can both be used for training and tuning systems. Clinical TempEval 2016 will also include the time normalization task (a.k.a. TIMEX3 value task) that unfortunately had to be dropped from last year's competition.

== Data ==

Clinical TempEval will use [the THYME corpus](https://github.com/stylerw/thymedata), which has been annotated using an extension of ISO TimeML developed by [the THYME project](http://thyme.healthnlp.org/) for the annotation of times, events and temporal relations in clinical notes. Normalized time expression values are additionally being annotated by [the PHEME project](http://www.pheme.eu/).

All notes that are part of the Clinical TempEval corpus have been de-identified as much as possible while still leaving the dates intact. To access the source texts for the Clinical TempEval corpus, participants must agree to handle the data appropriately, formalized in the requirement that participants must sign a data use agreement with the Mayo Clinic. To request a data use agreement, please follow [the instructions at the THYME website](http://thyme.healthnlp.org/#Getting_access_to_the_THYME_corpus_and_gold_...).

**Please apply for a data use agreement as soon as possible!** Though we have greatly streamlined the process since Clinical TempEval 2015, the process may still take some time.

== Task Organizers ==

Steven Bethard, The University of Alabama at Birmingham
Leon Derczynski, The University of Sheffield
Guergana Savova, Boston Children's Hospital and Harvard Medical School
James Pustejovsky, Brandeis University
Marc Verhagen, Brandeis University

== Important Dates ==

Trial data ready: June 30, 2015
Training data ready: August 30, 2015
Test data ready: December 15, 2015
Evaluation start: January 10, 2015
Evaluation end: January 31, 2015
Paper submission due: February 28, 2016 [TBC]
Paper reviews due: March 31, 2016 [TBC]
Camera ready due: April 30, 2016 [TBC]
SemEval workshop: Summer 2016

== Participants ==

Interested researchers should:

* Sign up on the Google Group at https://groups.google.com/d/forum/clinical-tempeval
* File for a data use agreement following the instructions at http://thyme.healthnlp.org/#Getting_access_to_the_THYME_corpus_and_gold_...
* Begin developing their systems with the trial data available from https://github.com/stylerw/thymedata/releases/tag/1.1.0%2Bcoloncancer-dev

If you have any further questions about the task, please post them to the Google Group where they will be answered by the organizers.