![[2508.12907] SNAP-UQ: Self-supervised Next-Activation Prediction for Single-Pass Uncertainty in TinyML](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2508.12907] SNAP-UQ: Self-supervised Next-Activation Prediction for Single-Pass Uncertainty in TinyML
Summary
The paper presents SNAP-UQ, a novel method for single-pass uncertainty estimation in TinyML, enhancing reliability in on-device monitoring without extensive resource demands.
Why It Matters
Reliable uncertainty estimation is crucial for TinyML applications, where resource constraints are significant. SNAP-UQ addresses the limitations of existing methods by providing an efficient, single-pass solution that maintains performance while minimizing resource use, making it highly relevant for developers working with constrained devices.
Key Takeaways
- SNAP-UQ enables single-pass uncertainty estimation without additional resource overhead.
- The method improves accuracy-drop detection and failure detection in TinyML applications.
- It reduces flash and latency requirements compared to traditional uncertainty estimation techniques.
- SNAP-UQ utilizes depth-wise next-activation prediction for more efficient monitoring.
- The approach is particularly beneficial for microcontrollers operating under strict resource constraints.
Computer Science > Machine Learning arXiv:2508.12907 (cs) [Submitted on 18 Aug 2025 (v1), last revised 18 Feb 2026 (this version, v4)] Title:SNAP-UQ: Self-supervised Next-Activation Prediction for Single-Pass Uncertainty in TinyML Authors:Ismail Lamaakal, Chaymae Yahyati, Khalid El Makkaoui, Ibrahim Ouahbi, Yassine Maleh View a PDF of the paper titled SNAP-UQ: Self-supervised Next-Activation Prediction for Single-Pass Uncertainty in TinyML, by Ismail Lamaakal and 4 other authors View PDF Abstract: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...
