The research objective of this CAREER proposal is to develop computer hardware architectures that would accelerate cognitive applications based on hierarchical neocortex models. These new class of cognitive models have shown significant promise in describing the functioning of the neocortex. Cognitive applications include perception, natural language comprehension, and cognitive reflection and are important in a large variety of domains such as national security, medicine, transportation, industry, and science. Large scale implementations will also enable neurobiologists to evaluate new models of the neocortex. In this work, biologically inspired architectures will be developed for large scale FPGA systems and embedded processing applications. Once these are developed, a model to predict the performance of the cognitive models on other architectures will be investigated. The research work will enable real-time implementations of cognitive applications at the large scale through FPGA clusters and at the embedded scale through a specialized architecture.
The educational objectives of this proposal are aimed at benefiting students from the K-12 to the university level. At the university level, an architecture visualization tool to enhance learning of architecture concepts will be developed. In addition, the research ideas developed form this work will be introduced into the curriculum through classroom projects. At the K-12 level, a set of lesson plans and learning tools about computer technology will be developed for elementary school students. Elementary school teachers will be trained in the use of the tools and lesson plans through a series of workshops. This work will be implemented in collaboration with the "Call Me Mister" program at Clemson University (this program is aimed at increasing the number of minority school teachers in South Carolina).
