sabrina
Position
Research Staff Member
Company
IBM Research
Location
Bethesda MD UNITED STATES
Bio

Research Scientist at IBM Watson (Yorktown Heights)


I am currently a Research Staff Member and the Health Informatics PIC chair of the IBM T.J. Research Center, leading the development of evidence-based behavioral insights generation analytics in the Center for Computational Health. Before working for the healthcare vertical at the department of science-driven solution, I worked on natural language processing-related applications such as social media trend analysis and meeting video summarization. My professional goal is to bring my business acumen and years of analytics training on the table to help translate real-world industry problems into services that can be illuminated with analytics and embedded into existing workflow. I am also interested in developing intelligent adaptive interfaces, which can curate exogenous data sources (beyond EHR, including enterprise data) and digest patient-generated data from wearables and novel biodevices and sensors. My background and experience enable me to consult, publish and patent avidly in the areas of analytics service design, computational linguistics, knowledge discovery and medical informatics. I have 20+ patents (US, non-US) and 40+ publications and had been selected as European Google Anita Borg Scholar.

Overarchingly, my research interest ties closely to the marriage of medical informatics, artificial intelligence and human computer interaction. with a focus on constructing behavioral models from noisy contexts and developing adaptive services and intelligent interfaces that help users synthesize information from very large data collections in their own contexts. My current research focus is on innovative approaches of computing personalization and incorporating personalization analytics into service design in mobile computing environments. My previous research concerns the task of discovering the implicit semantics in multimedia archives, using a learning framework to integrate multimodal features (and patterns) that are characteristics of the systematic differences in human conversations.