Showing 1–1 of 1 results for author: Cheng, T H
-
Membrane Potential Distribution Adjustment and Parametric Surrogate Gradient in Spiking Neural Networks
Authors:
Siqi Wang,
Tee Hiang Cheng,
Meng-Hiot Lim
Abstract:
As an emerging network model, spiking neural networks (SNNs) have aroused significant research attentions in recent years. However, the energy-efficient binary spikes do not augur well with gradient descent-based training approaches. Surrogate gradient (SG) strategy is investigated and applied to circumvent this issue and train SNNs from scratch. Due to the lack of well-recognized SG selection rul…
▽ More
As an emerging network model, spiking neural networks (SNNs) have aroused significant research attentions in recent years. However, the energy-efficient binary spikes do not augur well with gradient descent-based training approaches. Surrogate gradient (SG) strategy is investigated and applied to circumvent this issue and train SNNs from scratch. Due to the lack of well-recognized SG selection rule, most SGs are chosen intuitively. We propose the parametric surrogate gradient (PSG) method to iteratively update SG and eventually determine an optimal surrogate gradient parameter, which calibrates the shape of candidate SGs. In SNNs, neural potential distribution tends to deviate unpredictably due to quantization error. We evaluate such potential shift and propose methodology for potential distribution adjustment (PDA) to minimize the loss of undesired pre-activations. Experimental results demonstrate that the proposed methods can be readily integrated with backpropagation through time (BPTT) algorithm and help modulated SNNs to achieve state-of-the-art performance on both static and dynamic dataset with fewer timesteps.
△ Less
Submitted 26 April, 2023;
originally announced April 2023.
