October 1, 2018
Searching for health information is one of the most frequent online activities. Unfortunately, the results retrieved often are confusing and contain documents from unreliable sources. To make matters worse, search engines rely heavily on terms found in source documents for indexing, which leads to inconsistent use of vocabularies. Lack of “intellectual” control over indexing vocabularies makes the representation process brittle and vulnerable to manipulations and it impacts how users seek information. In this talk, I will discuss three projects that are exploring ways to improve health information seeking by drawing upon techniques from machine learning and human-information interaction: 1) automated vocabulary and concept generation system, 2) an interface that employs visualization to support navigation and retrieval of health related information, and 3) a machine-learning based consumer health information system that learns from interactions with its users and can support personalized delivery of health information. The talk will conclude with a discussion of privacy and security measures and a new architecture for supporting the online dissemination of sensitive health information.