AURA: Enabling Subject Matter Experts to Construct Declarative Knowledge Bases from Science Textbooks

Ken Barker, Vinay K. Chaudhri, Shaw Yi Chaw, Peter Clark, Daniel Hansch, Bonnie E. John, Sunil Mishra, John Pacheco, Bruce W. Porter, Aaron Spaulding, Moritz Weiten

Abstract: The long-term goal of Project Halo is to build an application called Digital Aristotle that can answer questions on a variety of science topics and provide user and domain appropriate explanations. As a near-term goal, we are focusing on enabling Subject Matter Experts (SMEs) to construct declarative knowledge bases (KBs) from 50 pages of a science textbook in the domains of Physics (Giancoli 2004), Chemistry (Brown et al. 2003) and Biology (Campbell et al. 2001) in a way that the system can answer questions similar to those on an Advanced Placement (AP) exam. We will demonstrate the current state of a system called AURA that we have been developing as a contributing technology toward the goal of Digital Aristotle. The innovative features of AURA are that it supports knowledge formulation for a mixture of textual and nontextual knowledge, and question formulation using an interactive dialog based on simplified English. The nontextual knowledge may contain tables, chemical reactions, and mathematical equations. In an extensive usability testing of AURA, we have established the basic viability of the approach.