Hybrid ANN-SNN Pipeline with Local Plasticity

2026-06-18

AI Summary

This work addresses the challenge of effectively transferring rich representations from pretrained artificial neural networks (ANNs) to spiking neural networks (SNNs) without relying on end-to-end backpropagation, thereby preserving biological plausibility. The authors propose a hybrid ANN-SNN pipeline that employs a pretrained EfficientNet as an encoder, whose activations are converted into spike trains via rate coding and subsequently fed into a CoLaNET spiking classifier trained exclusively with local plasticity rules. This approach achieves, for the first time, seamless integration of high-performance ANNs into SNNs using only biologically inspired local learning mechanisms. Evaluated on a 64-class ImageNet subset, the method attains 99.09% accuracy—comparable to conventional deep networks—while maintaining both computational efficiency and strong biological realism.

Abstract

This work proposes a hybrid ANN-SNN pipeline that effectively leverages the rich embeddings of pretrained artificial neural networks (ANNs) to enable high-performance spiking neural networks (SNNs). The architecture couples a pretrained EfficientNet encoder with a CoLaNET spiking classifier. We convert the encoder's activations into spike trains via rate-coding and train the subsequent SNN classifier using local, biologically inspired learning rules, bypassing end-to-end gradient propagation. This approach achieves 99.09% accuracy on a 64-class ImageNet benchmark, demonstrating performance on par with conventional deep networks. The work presents a biologically plausible and efficient framework for adapting powerful pretrained encoders to downstream spiking neural network tasks.

Problem

• ANN-SNN transfer
• spiking neural networks
• pretrained encoders
• biologically plausible learning
• rate coding

Innovation

• hybrid ANN-SNN
• local plasticity
• rate coding
• spiking neural networks
• pretrained encoder