Hardware Architectures for Flexible Component-Based Hybrid Cognitive Systems

Building cognitive systems that can interact with real world environments encompasses fields ranging from cognitive science, research artificial intelligence, robotics and control and many other research areas. While significant advances have been made in all of these arenas, the integration of these capabilities into functional cognitive systems remains a significant challenge. Our approach is premised upon a number of observations and principles. We observe that a general, unified theory of cognition has proven elusive. Moreover, existing monolithic system development approaches cannot efficiently implement and integrate the full spectrum of cognitive knowledge representation and algorithms, nor can they take advantage of novel computer architectures and capabilities. In the proposed Phase II effort, we will develop an end-to-end framework, methodology and supporting tools for cognitive systems development. The system-oriented development approach does not focus exclusively on single theories of cognition. Rather, it addresses the design and specification of components and the integration and optimization of components into coherent cognitive systems. It will include functional and behavioral components, control frameworks, and support tools and runtime infrastructures. The goal of this effort is to make the approach sufficiently complete and usable so that it can be used in research into cognitive architectures and systems.