Advancements in adapting deep convolution architectures for Spiking Neural Networks (SNNs) have significantly enhanced image classification performance and reduced computational burdens. However, the inability of Multiplication-Free Inference (MFI) to harmonize with attention and transformer mechanisms, which are critical to superior performance on high-resolution vision tasks, imposes limitations on these gains. To address this, our research explores a new pathway, drawing inspiration from the progress made in Multi-Layer Perceptrons (MLPs). We propose an innovative spiking MLP architecture that uses batch normalization to retain MFI compatibility and introduces a spiking patch encoding layer to reinforce local feature extraction capabilities. As a result, we establish an efficient multi-stage spiking MLP network that effectively blends global receptive fields with local feature extraction for comprehensive spike-based computation. Without relying on pre-training or sophisticated SNN training techniques, our network secures a top-1 accuracy of 66.39% on the ImageNet-1K dataset, surpassing the directly trained spiking ResNet-34 by 2.67%. Furthermore, we curtail computational costs, model capacity, and simulation steps. An expanded version of our network challenges the performance of the spiking VGG-16 network with a 71.64% top-1 accuracy, all while operating with a model capacity 2.1 times smaller. Our findings accentuate the potential of our deep SNN architecture in seamlessly integrating global and local learning abilities. Interestingly, the trained receptive field in our network mirrors the activity patterns of cortical cells.

翻译：脉冲神经网络（SNN）中深度卷积架构的适配进展显著提升了图像分类性能并降低了计算负担。然而，免乘推理（MFI）无法与对高分辨率视觉任务卓越性能至关重要的注意力机制及Transformer架构协同工作，这制约了上述收益。针对此问题，本研究借鉴多层感知器（MLP）的进展探索新路径。我们提出一种创新的尖峰MLP架构：采用批归一化保持MFI兼容性，并引入尖峰块编码层强化局部特征提取能力。由此构建的高效多阶段尖峰MLP网络，有效融合全局感受野与局部特征提取以实现全脉冲计算。无需预训练或复杂SNN训练技巧，该网络在ImageNet-1K数据集上取得66.39%的Top-1准确率，超越直接训练的尖峰ResNet-34达2.67%。此外，我们降低了计算开销、模型容量及仿真步数。扩展版网络以2.1倍更小的模型容量，以71.64%的Top-1准确率挑战尖峰VGG-16网络性能。我们的研究凸显了深度SNN架构在无缝整合全局与局部学习能力方面的潜力。值得关注的是，该网络训练所得的感受野与皮层细胞活动模式高度相似。