Research in Architectural Approaches to the Integration of Empirical, Analytic and Episodic Learning within SOAR

Abstract

Abstract : The goal of this research as originally stated: We propose to greatly expand the use and usefulness of the Soar architecture. Thus, our goal is to create a highly useable and available cognitive architecture that will greatly increase the use of cognitive architectures. To date, cognitive architectures have been either research or proprietary software. Soar has been one of the most successful cognitive architecture for developing knowledge-rich performance systems. We will greatly improve the development environment; develop a new, integrated debugger; expand and improve the tutorial and technical documentation; and provide training and support of the Soar architecture. All software will be freely available on SourceForge. During this project we have made significant progress in improving the Soar architecture and building a community. The major thrust of our work can be divided into three parts: 1. Evaluate the current state of Soar software, supporting documentation and tools. 2. Create a plan for future improvements to Soar and development of any new tools. 3. Make significant progress on the improvements to Soar and its associated tools and documentation.