Implementation of Delay-Enabled Axons for Lightweight Neural Networks with Improving Accuracy in Edge Computing
In this work, we constructed a novel axon with adjustable delay characteristics, which accurately emulates the delay property of biological axons. Based on this axon, we built a Liquid State Machine (LSM) network, which achieved a 5% increase in accuracy while reducing the number of parameters by a factor of 31.
Authors
- Peng Huang (ORCID: https://orcid.org/0000-0001-7954-8869)
- Hua Ai (ORCID: https://orcid.org/0009-0006-4007-0144)
- Lihao Zhang (ORCID: https://orcid.org/0000-0002-4746-0544)
- Zheng Zhou (ORCID: https://orcid.org/0009-0008-3957-3144)
- Guomeng Li (ORCID: https://orcid.org/0000-0001-9107-1522)
- Lifeng Liu (ORCID: https://orcid.org/0000-0002-0187-9243)
- Yu Feng (ORCID: https://orcid.org/0000-0001-7867-2788)
- Hongming Guan (ORCID: https://orcid.org/0000-0002-1466-4305)
Institutions
- Peking University (CN)
- Beijing Information Science & Technology University (CN)
Publication Details
- Published
- 2025-06-08
- DOI
- https://doi.org/10.23919/snw65111.2025.11097239
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00