RAN2SOM: a reconfigurable neural network architecture based on bit stream arithmetic

We introduce the RAN²SOM (Reconfigurable Architecture Neural Networks with Serially Operating Multipliers) architecture, a neural net architecture with a reconfigurable interconnection scheme based on bit stream arithmetic. RAN²SOM nets are implemented using field programmable gate array logic. By conducting the training phase in software and executing the actual application in hardware, conflicting demands can be met: training benefits from a fast edit-debug cycle, and once the design has stabilized a hardware implementation results in higher performance. While neural nets have been implemented in hardware in the past, larger digital nets have not been possible due to the real-estate requirements of single neutrons. We present a bit-serial encoding scheme and computation model, which allows space-efficient computation of the sum of weighted inputs, thereby facilitating the implementation of complex neural networks