A PyTorch-Compatible Spike Encoding Framework for Energy-Efficient Neuromorphic Applications

TL;DR

The paper addresses the challenge of using Spiking Neural Networks (SNNs) with traditional ML workflows by introducing a PyTorch-compatible, open-source spike-encoding framework that implements multiple temporal encoding algorithms (LIF, SF, PWM, BSA) and additional population and RL-oriented encoders. It provides a Converter-based design that encodes/decodes signals with optional hyperparameter optimization and evaluates performance on embedded hardware using a rigorous Optuna-driven parameter search. Empirical results across four synthetic signals show Step-Forward encoding ($SF$) generally offers the best energy efficiency and encoding speed while maintaining low reconstruction error, with other methods offering niche strengths depending on signal characteristics. The work delivers practical resources for deploying energy-efficient SNNs in ML and RL tasks on embedded platforms, including code, documentation, and a framework that facilitates integration into PyTorch-based pipelines.

Abstract

Spiking Neural Networks (SNNs) offer promising energy efficiency advantages, particularly when processing sparse spike trains. However, their incompatibility with traditional datasets, which consist of batches of input vectors rather than spike trains, necessitates the development of efficient encoding methods. This paper introduces a novel, open-source PyTorch-compatible Python framework for spike encoding, designed for neuromorphic applications in machine learning and reinforcement learning. The framework supports a range of encoding algorithms, including Leaky Integrate-and-Fire (LIF), Step Forward (SF), Pulse Width Modulation (PWM), and Ben's Spiker Algorithm (BSA), as well as specialized encoding strategies covering population coding and reinforcement learning scenarios. Furthermore, we investigate the performance trade-offs of each method on embedded hardware using C/C++ implementations, considering energy consumption, computation time, spike sparsity, and reconstruction accuracy. Our findings indicate that SF typically achieves the lowest reconstruction error and offers the highest energy efficiency and fastest encoding speed, achieving the second-best spike sparsity. At the same time, other methods demonstrate particular strengths depending on the signal characteristics. This framework and the accompanying empirical analysis provide valuable resources for selecting optimal encoding strategies for energy-efficient SNN applications.

Figures (7)

• Figure 1: SF Encoding Method
• Figure 2: LIF Encoding Method
• Figure 3: BSA Encoding Method
• Figure 4: PWM Encoding Method
• Figure 5: Four different signal types: (a) Vibration Signal, (b) Trended Signal, (c) Rectangular Signal, (d) Sinusoidal Signal.