Algorithm for Mapping Multilayer BP Networks Onto the SpiNNaker Neuromorphic Hardware

Provided by: Institute of Electrical & Electronic Engineers
Topic: Hardware
Format: PDF
In this paper the authors demonstrate the feasibility and evaluate the performance of using the SpiNNaker neuromorphic hardware to simulate traditional non-spiking multi-layer perception networks with the back propagation learning rule. In addition to investigating the mapping of checker-boarding partitioning scheme onto SpiNNaker, they propose a new algorithm called pipelined checker-boarding partitioning which introduces a pipelined mode and captures the parallelism within each partition of the weight matrix, allowing the overlapping of communication and computation.

Find By Topic