Abstract:

Many applications involve mining and analysis of various datasets. While more effective techniques exist for structured data, understanding and analyzing text involves extraction of entities and relationships. Furthermore, relating pieces of knowledge in terms of relevant entities and relationships from different documents (e.g., medical abstracts or publications, electronic medical records and clinical trial data) in a context specified by humans could help with unearthing Undiscovered Public Knowledge.

We have been carrying out fundamental research that utilizes Natural Language Processing and the emerging Semantic Web (also called Web 3.0) technologies for entity and relationship extraction from biomedical text. The proposed project focuses on development of a Semantic Browser, with initial focus on biomedical research. In this context, Computer Scientists at the Kno.e.sis Center of the Wright State University (WSU) will closely collaborate with Biomedical Researchers at the Computational Medicine Center (CMC) of the Cincinnati Children’s Medical Center (CCHMC), along with Angela Qu, an employee at P&G and a PhD student at CCHMC.

The primary outcome of this collaborative project will be the completion of our Semantic Browser prototype and evaluation of underlying entity and relationship extraction algorithms for biomedical text. If successful, this work will be lead to a key technology necessary for developing a comprehensive knowledge base that can maximally represent reusable knowledge components from pharmacological and biological domains, leverage comprehensive knowledge of principles of drug action and disease mechanisms to find sophisticated connections between drugs, diseases and individual patients. We anticipate that in the future, our ontology-driven semantic browsing and analysis can be applied to many other domains.

Older versions:
Semantic Browser Demo - V1 [AVI 29MB]