Abstract:Spiking Neural Networks (SNNs) have indeed shown remarkable promise in the field of computer vision, emerging as a low-energy alternative to traditional Artificial Neural Networks (ANNs). However, SNNs also face several challenges: i) Existing SNNs are not purely additive and involve a substantial amount of floating-point computations, which contradicts the original design intention of adapting to neuromorphic chips; ii) The incorrect positioning of convolutional and pooling layers relative to spiking layers leads to reduced accuracy; iii) Leaky Integrate-and-Fire (LIF) neurons have limited capability in representing local information, which is disadvantageous for downstream visual tasks like semantic segmentation. To address the challenges in SNNs, i) we introduce Pure Sparse Self Attention (PSSA) and Dynamic Spiking Membrane Shortcut (DSMS), combining them to tackle the issue of floating-point computations; ii) the Spiking Precise Gradient downsampling (SPG-down) method is proposed for accurate gradient transmission; iii) the Group-LIF neuron concept is introduced to ensure LIF neurons' capability in representing local information both horizontally and vertically, enhancing their applicability in semantic segmentation tasks. Ultimately, these three solutions are integrated into the Powerful Sparse-Spike-Driven Transformer (PSSD-Transformer), effectively handling semantic segmentation tasks and addressing the challenges inherent in SNNs. The experimental results demonstrate that our model outperforms previous results on standard classification datasets and also shows commendable performance on semantic segmentation datasets. Up to this point, PSSD is the first model in the SNN field to perform semantic segmentation on large datasets. The code will be made publicly available after the paper is accepted for publication.

ITERATED DILATED CONVOLUTIONAL NEURAL NETWORKS FOR WORD SEGMENTATION

PSSD-Transformer: Powerful Sparse Spike-Driven Transformer for Image Semantic Segmentation

Bi-directional LSTM Recurrent Neural Network for Chinese Word Segmentation

State-of-the-art Chinese Word Segmentation with Bi-LSTMs

ISDNet: Integrating Shallow and Deep Networks for Efficient Ultra-high Resolution Segmentation

Hyper-Gated Recurrent Neural Networks for Chinese Word Segmentation.

DAG-based Long Short-Term Memory for Neural Word Segmentation

Long Short-Term Memory Neural Networks for Chinese Word Segmentation.

A Reliable Distributed Convolutional Neural Network for Biology Image Segmentation.

A Gap-Based Framework for Chinese Word Segmentation Via Very Deep Convolutional Networks

Toward Fast and Accurate Neural Discourse Segmentation

Chinese Word Segmentation with Heterogeneous Graph Neural Network

Fast and Accurate Neural Word Segmentation for Chinese

Neural Word Segmentation Learning for Chinese

A Deep Convolutional Neural Model for Character-Based Chinese Word Segmentation

A Parallel Convolutional Network Based on Spiking Neural Systems

Effective Neural Solution for Multi-criteria Word Segmentation

Hybrid Dilated Convolution Network Using Attentive Kernels for Real-Time Semantic Segmentation

DSNet: A Novel Way to Use Atrous Convolutions in Semantic Segmentation

Deep Learning for Chinese Word Segmentation and POS Tagging.