Abstract

Neuromorphic computing is a relatively new discipline of computer science,where the principles of biological brain's computation and memory are used tocreate a new way of processing information, based on networks of spikingneurons. Those networks can be implemented as both analog and digitalimplementations, where for the latter, the Field Programmable Gate Arrays(FPGAs) are a frequent choice, due to their inherent flexibility, allowing theresearchers to easily design hardware neuromorphic architecture (NMAs).Moreover, digital NMAs show good promise in simulating various spiking neuralnetworks because of their inherent accuracy and resilience to noise, as opposedto analog implementations. This paper presents an overview of digital NMAsimplemented on FPGAs, with a goal of providing useful references to variousarchitectural design choices to the researchers interested in digitalneuromorphic systems. We present a taxonomy of NMAs that highlights groups ofdistinct architectural features, their advantages and disadvantages andidentify trends and predictions for the future of those architectures.