Papers

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

Why we picked it

这是脉冲 Transformer 量化的一篇新论文,把 4-bit 权重和膜电位压缩后仍能保住发放决策的准确性。做 SNN 边缘部署或事件相机任务(如 FE108 跟踪)的人可以直接参考它的校准思路,能耗降了 66.2% 还不掉点。

论文提出 Decision-Aligned Post-Training Quantization(DAQ),一种针对脉冲 Transformer 的后训练量化方法,在权重和膜电位变量均压缩到 4-bit 的情况下保持模型精度。

Body

其核心机制是将量化器校准到与浮点参考模型一致的神经元发放决策:论文分析了权重和膜电位量化误差如何在发放决策边界处产生影响,以及这些误差在循环神经元的时间步之间如何传播,并据此设计目标函数,让量化后的膜电位变量沿量化模型自身的状态轨迹保持在参考发放区域内。

实验覆盖多个脉冲 Transformer 架构,任务包括图像分类、语义分割和基于事件相机 的目标跟踪。在 FE108 数据集上,DAQ 配合 SDTrack 取得 58.8% AUC 和 90.9% PR,同时将能耗降低 66.2%,对做边缘部署和脉冲神经网络优化的研究者有直接参考价值。

Original · arXiv cs.AI

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

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\%.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include spiking neural networks, post-training quantization, Spiking Transformer, SDTrack.

Source attribution: curated from arXiv cs.AI (@Jialin Chen, Xiaochen Zhao, Aili Wang).