$$Events$$

Sep. 21, 2014
12:15

Room 201, Alon Building for Hi-Tech (#37)



Israel’s centralized health system is a treasure trove for medical informatics studies. Since most clinical notes are in Hebrew, taking advantage of this potential rich knowledge source requires specific techniques, and tools. The same tools can also help medical researchers tap the deluge of Hebrew health consumer texts available online in medical QA sites and forums. Both cases require Natural Language Processing methods for the practice of medical informatics or Data Science in Medical Hebrew.
In this workshop, we will review medical informatics methods combining NLP and dive in two use cases dealing with Hebrew medical texts. The workshop will include hands-on practice with the Medical Hebrew NLP software pipeline recently developed at BGU.

More adventurous participants are invited to remain for a hackathon over night, where the tools will be put in practice to develop practical software tools manipulating and mining Hebrew medical text.
 
 

Schedule:

12:15 - 12:30 Gathering / Coffee

12:30 - 12:35 Opening Remarks - Michael Elhadad

12:35 - 13:00 Introduction to Medical informatics and medical texts - [TBD]

13:00-14:30 Use Case I - Raphael Cohen
Classifying Epilepsy Status for Retrospective Study in Soroka. In depth study of how to solve a free text problem: methodology, baseline, NLP solutions and Hands-on lab using IPython.

14:30 - 14:45 Coffee Break

14:45-15:00 Use Case II - Yuval Barak Corren.
Identifying chief complain in ER admissions

15:00-15:15 Use Case III - Dmetry Leykin.
Identification of individual stress coping profile from social media

15:15-15:45 Understanding Medical Hebrew - Raphael Cohen. 
From tokenization to entity linking and syntactic parsing. We will present the pipeline for processing medical texts in Hebrew as well as the methods for dictionary / lexicon creation and approaches to text annotation.

15:45-15:55 Conclusion

15:55-16:10 Coffee Break

16:10-Midnight Hackathon - We will choose 1 or 2 use cases (depending on number of participants) divide into tasks and develop minimal code necessary to bootstrap work on the project

This event is free and open to the public. We kindly ask you to register, so the number of participants can be estimated for the required preparations.