摘要: |
The overarching objective of this effort is to provide a foundation for an affordable, ultra-dense, low-power computer system for processing spatio-temporal data on the fly. This includes construction of a mixed mode (including analog and digital circuits) neuromorphic computing system built for rapid configuration, dynamic adaptation, low-power operation, and that is well suited for processing spatio-temporal data. Neuromorphic or neuro-inspired computer architectures are particularly worthwhile given the increasing number of big data problems requiring techniques and systems that can capture knowledge from an abundance of data. Thus, the proposed memristor-based dynamic adaptive neural network array (mrDANNA) addresses contemporary application challenges while also enabling continued performance scaling. |