Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

UR2M: Uncertainty and Resource-Aware Event Detection on Microcontrollers

Hong Jia, Young D. Kwon, Dong Ma, Nhat Pham, Lorena Qendro, Tam Vu, Cecilia Mascolo

TL;DR

UR2M is presented, a novel Uncertainty and Resource-aware event detection framework for MCUs that develops an uncertainty-aware WED based on evidential theory for accurate event detection and reliable uncertainty estimation and introduces a cascade ML framework to achieve efficient model inference via early exits.

Abstract

Traditional machine learning techniques are prone to generating inaccurate predictions when confronted with shifts in the distribution of data between the training and testing phases. This vulnerability can lead to severe consequences, especially in applications such as mobile healthcare. Uncertainty estimation has the potential to mitigate this issue by assessing the reliability of a model's output. However, existing uncertainty estimation techniques often require substantial computational resources and memory, making them impractical for implementation on microcontrollers (MCUs). This limitation hinders the feasibility of many important on-device wearable event detection (WED) applications, such as heart attack detection. In this paper, we present UR2M, a novel Uncertainty and Resource-aware event detection framework for MCUs. Specifically, we (i) develop an uncertainty-aware WED based on evidential theory for accurate event detection and reliable uncertainty estimation; (ii) introduce a cascade ML framework to achieve efficient model inference via early exits, by sharing shallower model layers among different event models; (iii) optimize the deployment of the model and MCU library for system efficiency. We conducted extensive experiments and compared UR2M to traditional uncertainty baselines using three wearable datasets. Our results demonstrate that UR2M achieves up to 864% faster inference speed, 857% energy-saving for uncertainty estimation, 55% memory saving on two popular MCUs, and a 22% improvement in uncertainty quantification performance. UR2M can be deployed on a wide range of MCUs, significantly expanding real-time and reliable WED applications.

UR2M: Uncertainty and Resource-Aware Event Detection on Microcontrollers

TL;DR

UR2M is presented, a novel Uncertainty and Resource-aware event detection framework for MCUs that develops an uncertainty-aware WED based on evidential theory for accurate event detection and reliable uncertainty estimation and introduces a cascade ML framework to achieve efficient model inference via early exits.

Abstract

Traditional machine learning techniques are prone to generating inaccurate predictions when confronted with shifts in the distribution of data between the training and testing phases. This vulnerability can lead to severe consequences, especially in applications such as mobile healthcare. Uncertainty estimation has the potential to mitigate this issue by assessing the reliability of a model's output. However, existing uncertainty estimation techniques often require substantial computational resources and memory, making them impractical for implementation on microcontrollers (MCUs). This limitation hinders the feasibility of many important on-device wearable event detection (WED) applications, such as heart attack detection. In this paper, we present UR2M, a novel Uncertainty and Resource-aware event detection framework for MCUs. Specifically, we (i) develop an uncertainty-aware WED based on evidential theory for accurate event detection and reliable uncertainty estimation; (ii) introduce a cascade ML framework to achieve efficient model inference via early exits, by sharing shallower model layers among different event models; (iii) optimize the deployment of the model and MCU library for system efficiency. We conducted extensive experiments and compared UR2M to traditional uncertainty baselines using three wearable datasets. Our results demonstrate that UR2M achieves up to 864% faster inference speed, 857% energy-saving for uncertainty estimation, 55% memory saving on two popular MCUs, and a 22% improvement in uncertainty quantification performance. UR2M can be deployed on a wide range of MCUs, significantly expanding real-time and reliable WED applications.
Paper Structure (26 sections, 8 equations, 10 figures)

This paper contains 26 sections, 8 equations, 10 figures.

Table of Contents

  1. Introduction
  2. Related Works
  3. UR2M System overview
  4. Efficient Uncertainty Quantification
  5. Evidential Deep Learning
  6. Efficient Evidential Modeling for Event Detection on MCUs
  7. Uncertainty-aware training and optimization
  8. Cascade learning
  9. Single-event Sharing
  10. Multiple-event Sharing
  11. Implementation
  12. System Implementation
  13. Uncertainty Operator Implementation
  14. MCU Library Optimization
  15. Evaluation settings
  16. ...and 11 more sections

Figures (10)

  • Figure 1: Memory and power comparison between a typical mobile phone and microcontrollers.
  • Figure 2: System overview.
  • Figure 3: The Search and training of UR2M
  • Figure 4: Deployment stage. (a) Uncertainty deployment on MCU based on multiple operators to calculate uncertainty and classification results. (b) MCU library space before optimization (top) and after optimization (bottom).
  • Figure 5: Model sizes vs. accuracy and early exit result for single events. Note that the ECG5000 UB event has only one test sample.
  • ...and 5 more figures