FFGAF-SNN: The Forward-Forward Based Gradient Approximation Free Training Framework for Spiking Neural Networks
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908474775961600 |
|---|---|
| author | Xu, Changqing Yang, Ziqiang Liu, Yi Liao, Xinfang Mo, Guiqi Zeng, Hao Yang, Yintang |
| author_facet | Xu, Changqing Yang, Ziqiang Liu, Yi Liao, Xinfang Mo, Guiqi Zeng, Hao Yang, Yintang |
| contents | Spiking Neural Networks (SNNs) offer a biologically plausible framework for energy-efficient neuromorphic computing. However, it is a challenge to train SNNs due to their non-differentiability, efficiently. Existing gradient approximation approaches frequently sacrifice accuracy and face deployment limitations on edge devices due to the substantial computational requirements of backpropagation. To address these challenges, we propose a Forward-Forward (FF) based gradient approximation-free training framework for Spiking Neural Networks, which treats spiking activations as black-box modules, thereby eliminating the need for gradient approximation while significantly reducing computational complexity. Furthermore, we introduce a class-aware complexity adaptation mechanism that dynamically optimizes the loss function based on inter-class difficulty metrics, enabling efficient allocation of network resources across different categories. Experimental results demonstrate that our proposed training framework achieves test accuracies of 99.58%, 92.13%, and 75.64% on the MNIST, Fashion-MNIST, and CIFAR-10 datasets, respectively, surpassing all existing FF-based SNN approaches. Additionally, our proposed method exhibits significant advantages in terms of memory access and computational power consumption. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_23643 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FFGAF-SNN: The Forward-Forward Based Gradient Approximation Free Training Framework for Spiking Neural Networks Xu, Changqing Yang, Ziqiang Liu, Yi Liao, Xinfang Mo, Guiqi Zeng, Hao Yang, Yintang Computer Vision and Pattern Recognition Spiking Neural Networks (SNNs) offer a biologically plausible framework for energy-efficient neuromorphic computing. However, it is a challenge to train SNNs due to their non-differentiability, efficiently. Existing gradient approximation approaches frequently sacrifice accuracy and face deployment limitations on edge devices due to the substantial computational requirements of backpropagation. To address these challenges, we propose a Forward-Forward (FF) based gradient approximation-free training framework for Spiking Neural Networks, which treats spiking activations as black-box modules, thereby eliminating the need for gradient approximation while significantly reducing computational complexity. Furthermore, we introduce a class-aware complexity adaptation mechanism that dynamically optimizes the loss function based on inter-class difficulty metrics, enabling efficient allocation of network resources across different categories. Experimental results demonstrate that our proposed training framework achieves test accuracies of 99.58%, 92.13%, and 75.64% on the MNIST, Fashion-MNIST, and CIFAR-10 datasets, respectively, surpassing all existing FF-based SNN approaches. Additionally, our proposed method exhibits significant advantages in terms of memory access and computational power consumption. |
| title | FFGAF-SNN: The Forward-Forward Based Gradient Approximation Free Training Framework for Spiking Neural Networks |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2507.23643 |