ED-sKWS: Early-Decision Spiking Neural Networks for Rapid,and Energy-Efficient Keyword Spotting

Zeyang Song, Qianhui Liu, Qu Yang, Yizhou Peng, Haizhou Li
6/14/2024
cs.SDcs.AIeess.AS

No official code found. Best implementation: Pooookeman/ED-sKWS (12 stars, last commit 2024-11-22).

Abstract

Keyword Spotting (KWS) is essential in edge computing requiring rapid and energy-efficient responses. Spiking Neural Networks (SNNs) are well-suited for KWS for their efficiency and temporal capacity for speech. To further reduce the latency and energy consumption, this study introduces ED-sKWS, an SNN-based KWS model with an early-decision mechanism that can stop speech processing and output the result before the end of speech utterance. Furthermore, we introduce a Cumulative Temporal (CT) loss that can enhance prediction accuracy at both the intermediate and final timesteps. To evaluate early-decision performance, we present the SC-100 dataset including 100 speech commands with beginning and end timestamp annotation. Experiments on the Google Speech Commands v2 and our SC-100 datasets show that ED-sKWS maintains competitive accuracy with 61% timesteps and 52% energy consumption compared to SNN models without early-decision mechanism, ensuring rapid response and energy efficiency.

View on arXivView PDF

Code Implementations(1)

Pooookeman/ED-sKWSCommunity100%

[Interspeech 2024] Official implementation of "ED-sKWS: Early-Decision Spiking Neural Networks for Rapid,and Energy-Efficient Keyword Spotting"

120PythonNov 18, 20241 years ago

Similar Papers

Splitformer: An improved early-exit architecture for automatic speech recognition on edge devices

83% match1 repo62025

Training dynamic models using early exits for automatic speech recognition on resource-constrained devices

82% match1 repo62023

DBHN-Net: Dual-Branch Hybrid Neural Network For Low-Complexity Monaural Speech Enhancement

81% match2026

DAISY: Data Adaptive Self-Supervised Early Exit for Speech Representation Models

81% match1 repo22024

A Hybrid ANN-SNN Architecture for Low-Power and Low-Latency Visual Perception

81% match1 repo422023

Turbocharge Speech Understanding with Pilot Inference

80% match2023

Cite this paper

@article{song2024edskws,
  title  = {ED-sKWS: Early-Decision Spiking Neural Networks for Rapid,and Energy-Efficient Keyword Spotting},
  author = {Zeyang Song and Qianhui Liu and Qu Yang and Yizhou Peng and Haizhou Li},
  year   = {2024},
  eprint = {2406.12726},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2406.12726v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2406.12726.svg)](https://codeofpaper.com/paper/2406.12726)

Discussion