Discovering medical information from user-generated content

Abstract: Collecting medical information from large cohorts is expensive and is frequently biased by the difficulty people experience in reporting effects which have late onset, are due to several confounding effects, or are related to sensitive subjects. In this talk, I will show that specific types of User Generated Content (UGC) are less influenced by such biases, and are thus a low-cost alternative for extracting medical information from very large populations. I will demonstrate our approach to learning from UGC using examples from post-market drug surveillance, the stages of information seeking by cancer patients, and the detrimental results of good intention on anorexia patients.

תאריך: 
03/04/2013 - 15:00
מרצה: 
Elad Yom-Tov
דוא"ל להרשמה: 
Affiliation: 
Microsoft
מיקום: 
building 1103, Room 329