A Universal Abstract-Time Platform for Real-Time Neural Networks

Provided by: Institute of Electrical & Electronic Engineers
Topic: Hardware
Format: PDF
High-speed asynchronous hardware makes it possible to virtualized neural networks' temporal dynamics as well as their structure. Through SpiNNaker, a dedicated neural chip multiprocessor, the authors introduce a real-time modeling architecture that makes the neural model run on the device independent of the hardware specifics. The central features of this modeling architecture are: native concurrency, ability to support very large networks, and decoupling of the temporal and spatial characteristics of the model from those of the hardware.

Find By Topic