Defining and Organizing Medical NLP Tasks


Fri Sep 20, 2013


Ricky Kiyotaka Taira
University of California, Los Angeles


Natural Language Processing




In this talk, we touch on some of the major issues related to building high quality deep understanding medical NLP systems. Given that it is unlikely that any one group can singly develop such a system, it is important to understand how tasks can be factored while maintaining a global master design focused on the latest theories of language to insure the overall system can achieve maximum potential with flexible consideration of speed requirements. We present a potential framework that helps developers understand how various NLP-related efforts and knowledge sources can be integrated. The aspects considered include: 1) computational issues dealing with defining layers of intermediate semantic structures to reduce the dimensionality of the NLP problem; 2) algorithmic issues in which we discuss state-of-the-art procedures used to map between various levels of the hierarchy; and 3) implementation issues directed to software developers. The objective of this talk is to educate the audience to the various levels of semantic representation for language processing (e.g., word level concepts, ontological concepts, logical relations, logical frames, discourse structures, etc.). The master design presents an architecture for which diverse efforts and resources in medical NLP can be integrated in a principled way.



Dr. Ricky Taira obtained his Bachelor’s degree in electrical engineering in 1982, and went on to receive a PhD in biomedical physics in 1988 from UCLA. He is now a Professor in the Department of Radiological Sciences at UCLA’s David Geffen School of Medicine. He has previously held a faculty position at the University of Washington. Dr. Taira is a member of the UCLA Medical Imaging Informatics Group, where he is the lead scientist. His past research interests have included development of picture archive and communication systems (PACS), medical knowledge bases (the KMeD project), and currently, natural language processing (NLP) systems for of medical reports. He is also conducts research on formal representations for disease processes. Dr. Taira teaches the Medical Knowledge Representation and Imaging Informatics courses as part of the UCLA training program in imaging-based medical informatics.


IMPORTANT NOTICE: This ReadyTalk service includes a feature that allows audio and any documents and other materials exchanged or viewed during the session to be recorded.  By joining this session, you automatically consent to such recordings. If you do not consent to the recording, discuss your concerns with the meeting host prior to the start of the recording or do not join the session. Please note that any such recordings may be subject to discovery in the event of litigation.