提出单次推理的无标签不确定性估计方法，适配微型设备部署。
SNAP-UQ: Self-supervised Next-Activation Prediction for Single-Pass Uncertainty in TinyML
- 通过预测下层激活值实现单次前向传播的不确定性评估。
- 相比基线减少40%-60%闪存占用，提速25%-35%，内存超限问题少。
- 适用于资源受限的边缘设备，尤其适合实时故障检测场景。
在微型机器学习（TinyML）中，可靠的不确定性估计是设备端监控的关键缺失环节：微控制器必须在严格闪存与延迟预算下检测故障、分布偏移或准确率下降，但常见的不确定性方法（如深度集成、MC dropout、早期退出、时间缓冲）通常需要多次前向传播、额外分支或状态存储，难以在毫瓦级硬件上运行。本文提出一种新颖且实用的方法 SNAP-UQ，基于深度卷积层间激活预测实现单次前向传播、无需标签的不确定性估计。SNAP-UQ 选取少量主干网络层，使用极小的 int8 头，从前一层低秩投影中预测下一层激活的均值与尺度；标准化预测误差形成逐层意外信号，经轻量单调校准器聚合后输出可操作的不确定性分数。该设计不引入时间缓冲或辅助退出，保持无状态推理，仅增加几十千字节部署开销。在视觉与音频主干网络上，SNAP-UQ 相较于早期退出与深度集成基线，闪存减少约40%-60%，速度提升约25%-35%，多数竞争方法因内存超限无法部署。在受损数据流上，其准确性下降事件检测能力提升多个 AUPRC 点，单次前向传播下仍保持强故障检测性能（AUROC ≈ 0.9）。通过将不确定性建立在层间动态而非仅输出置信度之上，SNAP-UQ 为鲁棒的 TinyML 监控提供了新范式。代码已开源：https://github.com/Ism-ail11/SNAP-UQ
原文摘要 · Abstract (English)
Reliable uncertainty estimation is a key missing piece for on-device monitoring in TinyML: microcontrollers must detect failures, distribution shift, or accuracy drops under strict flash/latency budgets, yet common uncertainty approaches (deep ensembles, MC dropout, early exits, temporal buffering) typically require multiple passes, extra branches, or state that is impractical on milliwatt hardware. This paper proposes a novel and practical method, SNAP-UQ, for single-pass, label-free uncertainty estimation based on depth-wise next-activation prediction. SNAP-UQ taps a small set of backbone layers and uses tiny int8 heads to predict the mean and scale of the next activation from a low-rank projection of the previous one; the resulting standardized prediction error forms a depth-wise surprisal signal that is aggregated and mapped through a lightweight monotone calibrator into an actionable uncertainty score. The design introduces no temporal buffers or auxiliary exits and preserves state-free inference, while increasing deployment footprint by only a few tens of kilobytes. Across vision and audio backbones, SNAP-UQ reduces flash and latency relative to early-exit and deep-ensemble baselines (typically $\sim$40--60% smaller and $\sim$25--35% faster), with several competing methods at similar accuracy often exceeding MCU memory limits. On corrupted streams, it improves accuracy-drop event detection by multiple AUPRC points and maintains strong failure detection (AUROC $\approx 0.9$) in a single forward pass. By grounding uncertainty in layer-to-layer dynamics rather than solely in output confidence, SNAP-UQ offers a novel, resource-efficient basis for robust TinyML monitoring. Our code is available at: https://github.com/Ism-ail11/SNAP-UQ
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。