Principal Computational Linguist | MITRE
Dr. Karine Megerdoomian is a Principal Computational Linguist in the Human Language Technology Department at the MITRE Corporation, a Federally Funded Research and Development Center. She specializes in Artificial Intelligence, social media analytics, and lexical semantics. Karine has expertise in linguistically-informed computational approaches for less commonly studied languages, with a strong specialization on Middle Eastern languages. More recently, Karine’s work explores the applications of linguistic analysis and computational approaches in the legal domain. Karine has served on various conference and workshop program committees and has taught courses in Social Media Analytics and Narrative Analysis at Georgetown University.
The U.S. Probation and Pretrial Services Office (PPSO) staff supervise more than 300,000 people a year and produce billions of pages of information on defendants’ and offenders’ profile and conduct. While it is critical for probation officers to have up-to-date knowledge on their clients, the data are often stored in narrative texts in multiple large documents. As a result, these records remain mostly out of reach without the use of painstaking manual review. We developed an analytic prototype to automatically acquire structured information from natural language text in probation office documents through the application of PDF content extraction, text mining, and language analytics. Since serious mental illness is very prevalent in the U.S. corrections system, the first phase of the project focused on extracting information and constructing timelines from narrative text regarding the defendants’ mental health conditions, substance use, and treatment history. The prototype leverages rich linguistic and semantic information through the application of open-source Natural Language Processing systems, adapted for the existing use case by applying vector space semantics and machine learning techniques to enhance the results. To build a timeline of important events, we developed a temporal reasoning component able to identify relations between events, tightly integrated with the system’s parse and semantic dependency relations. Automated narrative extraction and the construction of an event timeline for defendants’ mental and emotional health history have allowed PPSO to have a better understanding of their client population and to perform analyses that were previously unavailable to the organization.