ReckOn is a spiking recurrent neural network (RNN) processor enabling on-chip learning over second-long timescales based on a modified version of the e-prop algorithm (we released a PyTorch implementation of the vanilla e-prop algorithm for leaky integrate-and-fire neurons here). It was prototyped and measured in 28-nm FDSOI CMOS at the Institute of Neuroinformatics, University of Zurich and ETH Zurich, and published at the 2022 IEEE International Solid-State Circuits Conference (ISSCC).
|ReckOn: A 28nm sub-mm² task-agnostic spiking recurrent neural network processor enabling on-chip learning over second-long timescales
|C. Frenkel and G. Indiveri
|IEEE International Solid-State Circuits Conference (ISSCC)