University of North Carolina and Covar Applied Technologies
ARAGORN will provide a flexible, and efficient question-answering pipeline. It will be based on a modular architecture for acquiring relevant data and composing cohesive answers to users’ biomedical questions. ARAGORN will include novel modules for: 1) enhancing/augmenting users’ questions; 2) finding and querying relevant KPs in an efficient manner; 3) scoring resulting answer-subgraphs by relevance/cohesion; and 4) coalescing answer subgraphs into concise and informative answers. These individual components will also be exposed independently to enable integration with other ARAs to complement and augment broader Translator capabilities.