Rethinking Attention Locality in Spiking Transformers

• URL: http://arxiv.org/abs/2608.08541v1
• Date: Sun, 09 Aug 2026 07:36:17 GMT
• Title: Rethinking Attention Locality in Spiking Transformers
• Abstract summary: Spiking Transformers provide a promising paradigm for efficient visual processing with spike-driven computation.<n>Their Softmax-free Spiking Self-Attention (SSA) struggles to establish spatially localized token interactions.<n>We propose Spatially Contiguous Local Attention with Boundary Continuity Pathway (SCLA-BCP)<n>Our approach improves mAP@50 by up to 9.50% on COCO on 2017 and mIoU by up to 3.42% on ADE20K.
• Score: 50.00501275379363
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Spiking Transformers provide a promising paradigm for efficient visual processing with spike-driven computation, yet their Softmax-free Spiking Self-Attention (SSA) struggles to establish spatially localized token interactions. Although existing locality-enhanced SSA methods improve accuracy, it remains unclear whether they consistently induce spatial locality across layers and different Spiking Transformer architectures. Through Mean Attention Distance (MAD) analysis, we reveal that computational locality does not necessarily translate into spatial locality and show that uniformly applying the same locality enhancement overlooks architecture-dependent deployment requirements. Motivated by these observations, we propose Spatially Contiguous Local Attention with Boundary Continuity Pathway (SCLA-BCP). SCLA computes attention within non-overlapping regions of spatially adjacent tokens, while BCP facilitates cross-boundary information exchange through a lightweight convolutional pathway. Furthermore, we develop a hierarchical locality deployment strategy to effectively apply SCLA-BCP across the two major Spiking Transformer architectures. Extensive experiments on seven static and neuromorphic datasets covering classification, detection, and segmentation demonstrate consistent improvements with limited parameter and energy overhead. Notably, our approach improves mAP@50 by up to 9.50% on COCO 2017 and mIoU by up to 3.42% on ADE20K. Visualizations, MAD analysis, and ablation studies further validate its effectiveness.

Related papers

• Spiking Local Interaction and Adaptive Complementary Fusion for Spiking Transformer [19.12698354620558]
  Spiking Transformers model token interactions primarily through spiking self-attention (SSA)<n>We introduce Spiking Local Interaction (SLI) and Adaptive Complementary Fusion (ACF)<n>ACF integrates SSA and SLI through layer-specific, channel-wise coefficients that adaptively balance their contributions at different network depths.
  arXiv  Detail & Related papers  (2026-08-11T14:23:26Z)
• Hierarchical Consistency Learning for Test-time Adaptation in Camouflage Perception [50.278200968044665]
  Camouflaged object detection (COD) aims to localize targets that exhibit minimal perceptual differences from backgrounds through physical attributes.<n>Existing methods, constrained by the static train-then-freeze paradigm, suffer from domain rigidity and dependency annotation.<n>We propose the hierarchical consistency learning framework, which integrates test-time adaptation for dynamic representation recalibration.
  arXiv  Detail & Related papers  (2026-05-25T09:57:46Z)
• RELO: Reinforcement Learning to Localize for Visual Object Tracking [70.27265738642956]
  We introduce RELO, a REinforcement-learning-to-LOcalize method for visual object tracking.<n>RELO replaces handcrafted spatial priors with a localization policy learned over spatial positions.<n>We show that RELO achieves 57.5% AUC on LaSOText without template updates.
  arXiv  Detail & Related papers  (2026-05-08T07:34:29Z)
• ViCrop-Det: Spatial Attention Entropy Guided Cropping for Training-Free Small-Object Detection [6.1221124436192875]
  ViCrop-Det is a training-free inference framework that introduces adaptive spatial trust region shrinkage.<n>We show that ViCrop-Det consistently adds +1-3 mAP@50 to RT-DETR-R50 and Deformable DETR with a marginal 20-23% latency overhead.<n>Under compute-matched settings, our adaptive routing strategy comprehensively surpasses uniform slicing baselines, achieving a highly optimized accuracy-speed trade-off.
  arXiv  Detail & Related papers  (2026-04-29T15:35:03Z)
• OVS-DINO: Open-Vocabulary Segmentation via Structure-Aligned SAM-DINO with Language Guidance [53.45125687428705]
  Open-Vocabulary aims to segment image regions beyond predefined category sets by leveraging semantic descriptions.<n>Recent efforts have incorporated Vision Foundation Models (VFMs) like DINO to alleviate these limitations.<n>We propose OVS-DINO, a novel framework that revitalizes latent edge-sensitivity of DINO through structural alignment.
  arXiv  Detail & Related papers  (2026-04-09T16:57:11Z)
• OCTOPUS: Enhancing the Spatial-Awareness of Vision SSMs with Multi-Dimensional Scans and Traversal Selection [20.717476762904038]
  We introduce OCTOPUS, a novel architecture that preserves both global context and local spatial structure within images.<n>OCTOPUS performs discrete reoccurrence along eight principal orientations, going forward or backward in the horizontal, vertical, and diagonal directions.<n>In our classification and segmentation benchmarks, OCTOPUS demonstrates notable improvements in boundary preservation and region consistency.
  arXiv  Detail & Related papers  (2026-01-31T21:12:59Z)
• Cross-Scan Mamba with Masked Training for Robust Spectral Imaging [51.557804095896174]
  We propose the Cross-Scanning Mamba, named CS-Mamba, that employs a Spatial-Spectral SSM for global-local balanced context encoding.<n>Experiment results show that our CS-Mamba achieves state-of-the-art performance and the masked training method can better reconstruct smooth features to improve the visual quality.
  arXiv  Detail & Related papers  (2024-08-01T15:14:10Z)
• Exploring Consistency in Cross-Domain Transformer for Domain Adaptive Semantic Segmentation [51.10389829070684]
  Domain gap can cause discrepancies in self-attention. Due to this gap, the transformer attends to spurious regions or pixels, which deteriorates accuracy on the target domain. We propose adaptation on attention maps with cross-domain attention layers.
  arXiv  Detail & Related papers  (2022-11-27T02:40:33Z)
• Weakly Supervised Object Localization via Transformer with Implicit Spatial Calibration [20.322494442959762]
  Weakly Supervised Object Localization (WSOL) has attracted much attention because of its low annotation cost in real applications. We introduce a simple yet effective Spatial Module (SCM) for accurate WSOL, incorporating semantic similarities of patch tokens and their spatial relationships into a unified diffusion model. SCM is designed as an external module of Transformer, and can be removed during inference to reduce the computation cost.
  arXiv  Detail & Related papers  (2022-07-21T12:37:15Z)
• An Entropy-guided Reinforced Partial Convolutional Network for Zero-Shot Learning [77.72330187258498]
  We propose a novel Entropy-guided Reinforced Partial Convolutional Network (ERPCNet) ERPCNet extracts and aggregates localities based on semantic relevance and visual correlations without human-annotated regions. It not only discovers global-cooperative localities dynamically but also converges faster for policy gradient optimization.
  arXiv  Detail & Related papers  (2021-11-03T11:13:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.