arXiv：cs.AI· Jingang Jin, Yue Ma, Zhenhang Zhang, Qinru Qiu·· 16 小时前
基于 FPGA 的脉冲神经网络实现时空模式学习与分类
An FPGA-Based Spiking Neural Network for Spatiotemporal Pattern Learning and Classification
AI 导读
研究团队提出一种 FPGA 实现的多层脉冲神经网络（SNN），在同一硬件上完成推理与训练，采用时间前向训练算法规避 BPTT 的内存与延迟开销。在 MNIST 及五个多变量时间序列分类数据集上，全片上训练的 SNN 精度与使用 BPTT 训练的 SNN 或 LSTM 相当，相比 CPU 和 GPU 实现吞吐量提升 2.5 至 6 倍，每样本能耗低一个数量级。
正文
Abstract:Spiking Neural Networks (SNNs) are promising for energy-efficient spatiotemporal pattern recognition at the edge, but efficient deployment and on-device adaptation require specialized hardware. In this work, we present an FPGA implementation of multilayer SNNs for spatiotemporal pattern classification, in which both inference and training take place on the same hardware. Our design adopts a temporal-forward-only training algorithm that avoids the memory and latency overhead associated with traditional Backpropagation Through Time (BPTT). A mixed-precision scheme balances training stability with resource efficiency, and architectural optimization maximizes hardware and memory utilization. We further identify and resolve hazards that arise when inference and learning share the same memory, including port conflicts between the forward and backward access patterns and read-after-write dependencies in floating-point potential accumulation. Experimental evaluations on MNIST and five multivariate time-series classification datasets demonstrate that, with fully on-chip training, our SNN model achieves accuracy comparable to SNN or LSTM trained with BPTT. Compared to CPU and GPU based implementations, our FPGA design delivers 2.5 to 6 times higher throughput while consuming an order of magnitude less energy per sample.
| Comments: | 6 pages |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.17463 [cs.AI] |
| (or arXiv:2610.17463v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.17463 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jingang Jin [view email]
[v1]
Mon, 12 Oct 2026 22:26:18 UTC (369 KB)
来源：arXiv：cs.AI · arxiv.org