Energy Efficient Training of SNN using Local Zeroth Order Method
This work addresses energy-efficient training of SNNs for neuromorphic computing applications, offering an incremental improvement over existing surrogate methods.
The paper tackles the problem of training spiking neural networks (SNNs) by addressing gradient loss and non-differentiability due to the Heaviside function, proposing a local zeroth-order method that achieves a 100x speed-up in backward computation time and better generalization compared to state-of-the-art energy-efficient techniques.
Spiking neural networks are becoming increasingly popular for their low energy requirement in real-world tasks with accuracy comparable to the traditional ANNs. SNN training algorithms face the loss of gradient information and non-differentiability due to the Heaviside function in minimizing the model loss over model parameters. To circumvent the problem surrogate method uses a differentiable approximation of the Heaviside in the backward pass, while the forward pass uses the Heaviside as the spiking function. We propose to use the zeroth order technique at the neuron level to resolve this dichotomy and use it within the automatic differentiation tool. As a result, we establish a theoretical connection between the proposed local zeroth-order technique and the existing surrogate methods and vice-versa. The proposed method naturally lends itself to energy-efficient training of SNNs on GPUs. Experimental results with neuromorphic datasets show that such implementation requires less than 1 percent neurons to be active in the backward pass, resulting in a 100x speed-up in the backward computation time. Our method offers better generalization compared to the state-of-the-art energy-efficient technique while maintaining similar efficiency.