A Spike in Performance: Training Hybrid-Spiking Neural Networks with Quantized Activation Functions
arXiv:2002.03553
Abstract
The machine learning community has become increasingly interested in the energy efficiency of neural networks. The Spiking Neural Network (SNN) is a promising approach to energy-efficient computing, since its activation levels are quantized into temporally sparse, one-bit values (i.e., "spike" events), which additionally converts the sum over weight-activity products into a simple addition of weights (one weight for each spike). However, the goal of maintaining state-of-the-art (SotA) accuracy when converting a non-spiking network into an SNN has remained an elusive challenge, primarily due to spikes having only a single bit of precision. Adopting tools from signal processing, we cast neural activation functions as quantizers with temporally-diffused error, and then train networks while smoothly interpolating between the non-spiking and spiking regimes. We apply this technique to the Legendre Memory Unit (LMU) to obtain the first known example of a hybrid SNN outperforming SotA recurrent architectures -- including the LSTM, GRU, and NRU -- in accuracy, while reducing activities to at most 3.74 bits on average with 1.26 significant bits multiplying each weight. We discuss how these methods can significantly improve the energy efficiency of neural networks.
arXiv comment8 pages, 7 page supplementary
References in corpus (13)
- Adam: A Method for Stochastic Optimization
- To prune, or not to prune: exploring the efficacy of pruning for model compression
- A Simple Way to Initialize Recurrent Networks of Rectified Linear Units
- Hello Edge: Keyword Spotting on Microcontrollers
- Spiking Deep Networks with LIF Neurons
- Channel Gating Neural Networks
- Cold Case: The Lost MNIST Digits
- Discovering Low-Precision Networks Close to Full-Precision Networks for Efficient Embedded Inference
- Bit Efficient Quantization for Deep Neural Networks
- Energy-Aware Neural Architecture Optimization with Fast Splitting Steepest Descent
- Recognizing Images with at most one Spike per Neuron
- Towards Non-saturating Recurrent Units for Modelling Long-term Dependencies
- Improving Efficiency in Neural Network Accelerator Using Operands Hamming Distance optimization