Online Continual Learning on Intel Loihi 2 via a Co-designed Spiking Neural Network
CLP-SNN is a spiking neural network designed for online continual learning using Intel Loihi 2 hardware. It adapts the Continually Learning Prototypes (CLP) algorithm by replacing complex global updates with event-driven, local rules. This co-design achieves massive improvements in efficiency, delivering 113x lower latency and 6,600x lower energy compared to standard edge GPU methods while maintaining high accuracy.