EUVIMEDEUVIMED entdecken
Crossref

Fiber Transistors as a Hardware Surrogate Gradient for Backpropagation in Spiking Neural Networks

Weichu Chen iD, Hao Jiang iD, Chengyang Du iD, Yueheng Zhong iD, Xiangyu Wang iD, Xiang Li iD, Zhu Chen iD, Qicheng Liang iD, Fengqiang Sun iD, Yuwen Zhu iD, Jian-Gang Chen iD, Liang-Wen Feng iD und 4 weitere

Advanced Materials · 2026-01 · 38 (6)

DOI: 10.1002/adma.202514904 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Advanced Materials, online erschienen am 06.11.2025. Verfasst von Weichu Chen und 15 weiteren Autor:innen. Beteiligte Einrichtung: State Key Laboratory of Advanced Fiber Materials, College of Materials Science and Engineering, Donghua University, Shanghai, 201620, China. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 54 Literaturangaben. Gefördert von: National Key R&D Program of China, DHU Distinguished Young Professor Program, Taishan Industrial Experts Programme Special Funding.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Efficient training of spiking neural networks (SNNs) in flexible neuromorphic hardware remains a major challenge due to the non‐differentiable nature of spiking activations and the limited availability of trainable, energy‐efficient device platforms. Here, a tunable textile‐based vertical organic electrochemical transistor (TT‐vOECT) enables implementation of surrogate gradients computation for backpropagation in SNNs. Using a solvent interdiffusion solidification spinning strategy, Plateau–Rayleigh instabilities are overcome to fabricate coaxial trilayer fibers with well‐defined heterojunctions between PEDOT:PSS and BBL. The resulting TT‐vOECT exhibits nonlinear transfer characteristics that closely approximate the Sigmoid derivative. A conditionally activated backpropagation (CAB) mechanism is further proposed, in which synaptic updates are gated by both surrogate gradient magnitude and input spike, is implemented using reconfigurable logic arrays based on dual TT‐vOECTs. This framework enables sparse, event‐driven weight updates with reduced computational overhead. Integrated into a convolutional SNN, the device‐driven CAB strategy achieves high‐accuracy classification of electroencephalogram signals for multiclass neurological disorder diagnosis, comparable to software baselines, while reducing computational redundancy by ≈20%. The results establish a scalable and flexible organic hardware platform for trainable neuromorphic systems with biologically inspired learning dynamics.

Schlagwörter

Stichwörter der Autor:innen

EEG‐based neurological disorder classificationflexible organic electrochemical transistorsneuromorphic computingsolvent interdiffusion solidification spinningsurrogate gradient backpropagation

Autor:innen und Institutionen

  • State Key Laboratory of Advanced Fiber Materials, College of Materials Science and Engineering, Donghua University, Shanghai, 201620, China.
  • Shanghai Key Laboratory of Multidimensional Information Processing, School of Communication & Electronic Engineering, East China Normal University, Shanghai, 200241, China.
  • Key Laboratory of Green Chemistry & Technology, Ministry of Education, College of Chemistry, Sichuan University, Chengdu, 610065, China.
  • Henan Academy of Sciences, Zhengzhou, 450046, China.
  • Yuyue Home Textile Co., Ltd., Shandong, 256623, China.

ORCID-Profile: Weichu Chen, Hao Jiang, Chengyang Du, Yueheng Zhong, Xiangyu Wang, Xiang Li, Zhu Chen, Qicheng Liang, Fengqiang Sun, Yuwen Zhu, Jian-Gang Chen, Liang-Wen Feng, Hongzhi Wang, Meifang Zhu, Hengda Sun, Gang Wang

Förderung

  • National Key R&D Program of China
  • DHU Distinguished Young Professor Program
  • Taishan Industrial Experts Programme Special Funding
  • Key R&D Program of Shandong Province
  • Pujiang Project of the Shanghai Magnolia Talent Program
  • Sichuan Science and Technology Program
  • Science and Technology Commission of Shanghai Municipality
  • National Natural Science Foundation of China

Zitieren

APA

Chen, W., Jiang, H., Du, C., Zhong, Y., Wang, X., Li, X., Chen, Z., Liang, Q., Sun, F., Zhu, Y., Chen, J. G., Feng, L. W., Wang, H., Zhu, M., Sun, H., & Wang, G. (2026). Fiber Transistors as a Hardware Surrogate Gradient for Backpropagation in Spiking Neural Networks. Advanced Materials, 38(6). https://doi.org/10.1002/adma.202514904

Vancouver

Chen W, Jiang H, Du C, Zhong Y, Wang X, Li X, et al. Fiber Transistors as a Hardware Surrogate Gradient for Backpropagation in Spiking Neural Networks. Advanced Materials. 2026;38(6). doi:10.1002/adma.202514904

BibTeX
@article{Chen2026,
  author = {Chen, Weichu and Jiang, Hao and Du, Chengyang and Zhong, Yueheng and Wang, Xiangyu and Li, Xiang and Chen, Zhu and Liang, Qicheng and Sun, Fengqiang and Zhu, Yuwen and Chen, Jian-Gang and Feng, Liang-Wen and Wang, Hongzhi and Zhu, Meifang and Sun, Hengda and Wang, Gang},
  title = {Fiber Transistors as a Hardware Surrogate Gradient for Backpropagation in Spiking Neural Networks},
  journal = {Advanced Materials},
  year = {2026},
  volume = {38},
  number = {6},
  doi = {10.1002/adma.202514904}
}
Kostenlos starten →