跳到正文
arXiv cs.AI· Jialin Chen, Xiaochen Zhao, Aili Wang·· 2 小时前AI 评分35

DAQ:面向脉冲 Transformer 的决策对齐后训练量化方法

DAQ: Decision-Aligned Post-Training Quantization for Spiking Transformers

AI 导读

论文提出决策对齐后训练量化方法 DAQ,通过校准量化器使脉冲神经元的发放决策与浮点参考模型对齐,分析决策边界误差及其在时间步上的传播。该方法支持 4-bit 权重与膜变量量化,在图像分类、语义分割和事件跟踪上验证有效,在 FE108 上配合 SDTrack 达到 58.8% AUC 和 90.9% PR,同时降低 66.2% 能耗。

正文 · 原文

View PDF HTML (experimental)

Abstract:Spiking transformers combine attention-based representation learning with sparse spike communication, yet their weights and membrane variables still incur substantial storage and computation overhead. Post-training quantization (PTQ) offers a practical way to reduce these overheads, but preserving neuronal firing under weight and membrane quantization remains challenging. In this paper, we propose Decision-Aligned Post-Training Quantization (DAQ), a method that calibrates the quantizers to align neuronal firing decisions with those of a floating-point reference. We analyze how errors from these quantizers affect firing at decision boundaries and how they propagate over time steps in recurrent neurons. Based on this analysis, we formulate an objective that encourages quantized membrane variables to remain within the reference firing regions along the quantized model's own state trajectory. Experiments across multiple spiking transformer architectures on image classification, semantic segmentation, and event-based tracking demonstrate the effectiveness of DAQ with 4-bit weights and membrane variables. On FE108, DAQ achieves 58.8\% AUC and 90.9\% PR with SDTrack, while reducing energy by 66.2\%.
Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2611.22611 [cs.AI]
  (or arXiv:2611.22611v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.22611

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Xiaochen Zhao [view email]
[v1] Tue, 17 Nov 2026 11:25:44 UTC (3,872 KB)

来源:arXiv cs.AI · arxiv.org