发表机构
Humboldt University of Berlin; Tallinn University of Technology; Brandenburgische Technische Universität Cottbus-Senftenberg（柏林洪堡大学; 塔林理工大学; 科特布斯-森夫滕贝格勃兰登堡工业大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
BLADE 提出首个可靠性感知的 SNN-ANN 边界选择方法，联合优化可靠性、精度、时延与能耗，实现事件驱动目标检测的高能效可靠部署。
AI 中文摘要
混合脉冲神经网络（SNN）-人工神经网络（ANN）架构结合了 SNN 的能效优势与 ANN 在事件驱动目标检测中的卓越检测精度。然而，现有的混合 SNN-ANN 网络采用静态推理，并主要根据精度和能耗选择 SNN-ANN 边界，未考虑动态推理或可靠性。本文提出 BLADE，这是首个面向具有 ANN 早退机制的动态混合 SNN-ANN 网络的可靠性感知边界选择方法。所提出的框架根据可靠性、检测精度、执行时间和能耗联合优化 SNN-ANN 边界和 ANN 早退配置，同时在设计空间探索期间通过分层统计故障注入来纳入可靠性。在事件驱动目标检测器上的实验评估实现了 0.691 的 mAP 0.5，同时当 ANN 早退触发时将推理计算能耗降低至 15.82 mJ。可靠性分析确定最显著的浮点指数位是灾难性故障的主要来源，在其 58.8% 的故障注入中产生显著或更差的精度退化。以约 3% 的存储开销保护该单个位，可在评估的现实技术故障率下消除灾难性故障。此外，增加 SNN 计算的比例可提高容错能力，全 SNN 配置在激进故障条件下实现了 0.965 的可靠性保持。结果表明，联合优化可靠性、精度、执行时间和能耗，能够为安全关键的边缘 AI 应用实现更可靠的动态混合 SNN-ANN 系统部署。
英文摘要
Hybrid Spiking Neural Network (SNN)-Artificial Neural Network (ANN) architectures combine the energy efficiency of SNNs with the superior detection accuracy of ANNs for event-based object detection. Existing hybrid SNN--ANN networks, however, employ static inference and select the SNN-ANN boundary primarily according to accuracy and energy consumption, without considering dynamic inference or reliability. This paper presents BLADE, the first reliability-aware boundary selection methodology for dynamic hybrid SNN-ANN networks with ANN early exit. The proposed framework jointly optimizes the SNN-ANN boundary and ANN early-exit configuration according to reliability, detection accuracy, execution time, and energy consumption, while incorporating reliability through hierarchical statistical fault injection during design-space exploration. Experimental evaluation on an event-based object detector achieves an mAP 0.5 of 0.691 while reducing the inference compute energy to 15.82~mJ when the ANN early exit fires. Reliability analysis identifies the most significant floating-point exponent bit as the dominant source of catastrophic failures, producing significant-or-worse accuracy degradation in 58.8% of its fault injections. Protecting this single bit with approximately 3% storage overhead eliminates catastrophic failures across the evaluated realistic technology fault rates. Furthermore, increasing the proportion of SNN computation improves fault tolerance, with the fully SNN configuration achieving a reliability retention of 0.965 under aggressive fault conditions. The results demonstrate that jointly optimizing reliability, accuracy, execution time, and energy consumption enables more dependable deployment of dynamic hybrid SNN--ANN systems for safety-critical edge AI applications.
