Abstract:The intrinsic dynamics and event-driven nature of spiking neural networks (SNNs) make them excel in processing temporal information by naturally utilizing embedded time sequences as time steps. Recent studies adopting this approach have demonstrated SNNs' effectiveness in speech command recognition, achieving high performance by employing large time steps for long time sequences. However, the large time steps lead to increased deployment burdens for edge computing applications. Thus, it is important to balance high performance and low energy consumption when detecting temporal patterns in edge devices. Our solution comprises two key components. 1). We propose a high-performance fully spike-driven framework termed SpikeSCR, characterized by a global-local hybrid structure for efficient representation learning, which exhibits long-term learning capabilities with extended time steps. 2). To further fully embrace low energy consumption, we propose an effective knowledge distillation method based on curriculum learning (KDCL), where valuable representations learned from the easy curriculum are progressively transferred to the hard curriculum with minor loss, striking a trade-off between power efficiency and high performance. We evaluate our method on three benchmark datasets: the Spiking Heidelberg Dataset (SHD), the Spiking Speech Commands (SSC), and the Google Speech Commands (GSC) V2. Our experimental results demonstrate that SpikeSCR outperforms current state-of-the-art (SOTA) methods across these three datasets with the same time steps. Furthermore, by executing KDCL, we reduce the number of time steps by 60% and decrease energy consumption by 54.8% while maintaining comparable performance to recent SOTA results. Therefore, this work offers valuable insights for tackling temporal processing challenges with long time sequences in edge neuromorphic computing systems.

Time-Sensitive Semantic Communication Using Dynamic Spiking Neural Networks

RSNN: Recurrent Spiking Neural Networks for Dynamic Spatial-Temporal Information Processing

Deep CovDenseSNN: A Hierarchical Event-Driven Dynamic Framework with Spiking Neurons in Noisy Environment

SNN-SC: A Spiking Semantic Communication Framework for Collaborative Intelligence

DPSNN: Spiking Neural Network for Low-Latency Streaming Speech Enhancement

PSSD-Transformer: Powerful Sparse Spike-Driven Transformer for Image Semantic Segmentation

Towards Energy-Preserving Natural Language Understanding with Spiking Neural Networks

CSNN: an Augmented Spiking Based Framework with Perceptron-Inception

Deep Learning-Based Image Semantic Communication System

Ultra-low Latency Spiking Neural Networks with Spatio-Temporal Compression and Synaptic Convolutional Block

DynSNN: A Dynamic Approach to Reduce Redundancy in Spiking Neural Networks

Temporal-Coded Spiking Neural Networks with Dynamic Firing Threshold: Learning with Event-Driven Backpropagation

DRL-Driven Dynamic Resource Allocation for Task-Oriented Semantic Communication

Accurate and Efficient Event-based Semantic Segmentation Using Adaptive Spiking Encoder-Decoder Network

Deep Learning Enabled Task-Oriented Semantic Communication for Memory-Limited Devices

Delay learning based on temporal coding in Spiking Neural Networks

SpikingSSMs: Learning Long Sequences with Sparse and Parallel Spiking State Space Models

Intelligent Semantic Communication System Based on Kolmogorov–Arnold Networks Driven by Dynamic Terminal-Side Computing Power Network

Ternary Spike-based Neuromorphic Signal Processing System

Efficient Speech Command Recognition Leveraging Spiking Neural Network and Curriculum Learning-based Knowledge Distillation

PT-Spike: A Precise-Time-Dependent Single Spike Neuromorphic Architecture with Efficient Supervised Learning