Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MO-SAE:Multi-Objective Stacked Autoencoders Optimization for Edge Anomaly Detection

Lizhao Zhang, Shengsong Kong, Tao Guo, Shaobo Li, Zhenzhou Ji

Abstract

Stacked AutoEncoders (SAE) have been widely adopted in edge anomaly detection scenarios. However, the resource-intensive nature of SAE can pose significant challenges for edge devices, which are typically resource-constrained and must adapt rapidly to dynamic and changing conditions. Optimizing SAE to meet the heterogeneous demands of real-world deployment scenarios, including high performance under constrained storage, low power consumption, fast inference, and efficient model updates, remains a substantial challenge. To address this, we propose an integrated optimization framework that jointly considers these critical factors to achieve balanced and adaptive system-level optimization. Specifically, we formulate SAE optimization for edge anomaly detection as a multi-objective optimization problem and propose MO-SAE (Multi-Objective Stacked AutoEncoders). The multiple objectives are addressed by integrating model clipping, multi-branch exit design, and a matrix approximation technique. In addition, a multi-objective heuristic algorithm is employed to effectively balance the competing objectives in SAE optimization. Our results demonstrate that the proposed MO-SAE delivers substantial improvements over the original approach. On the x86 architecture, it reduces storage space and power consumption by at least 50%, improves runtime efficiency by no less than 28%, and achieves an 11.8% compression rate, all while maintaining application performance. Furthermore, MO-SAE runs efficiently on edge devices with ARM architecture. Experimental results show a 15% improvement in inference speed, facilitating efficient deployment in cloud-edge collaborative anomaly detection systems.

MO-SAE:Multi-Objective Stacked Autoencoders Optimization for Edge Anomaly Detection

Abstract

Stacked AutoEncoders (SAE) have been widely adopted in edge anomaly detection scenarios. However, the resource-intensive nature of SAE can pose significant challenges for edge devices, which are typically resource-constrained and must adapt rapidly to dynamic and changing conditions. Optimizing SAE to meet the heterogeneous demands of real-world deployment scenarios, including high performance under constrained storage, low power consumption, fast inference, and efficient model updates, remains a substantial challenge. To address this, we propose an integrated optimization framework that jointly considers these critical factors to achieve balanced and adaptive system-level optimization. Specifically, we formulate SAE optimization for edge anomaly detection as a multi-objective optimization problem and propose MO-SAE (Multi-Objective Stacked AutoEncoders). The multiple objectives are addressed by integrating model clipping, multi-branch exit design, and a matrix approximation technique. In addition, a multi-objective heuristic algorithm is employed to effectively balance the competing objectives in SAE optimization. Our results demonstrate that the proposed MO-SAE delivers substantial improvements over the original approach. On the x86 architecture, it reduces storage space and power consumption by at least 50%, improves runtime efficiency by no less than 28%, and achieves an 11.8% compression rate, all while maintaining application performance. Furthermore, MO-SAE runs efficiently on edge devices with ARM architecture. Experimental results show a 15% improvement in inference speed, facilitating efficient deployment in cloud-edge collaborative anomaly detection systems.
Paper Structure (12 sections, 8 equations, 6 figures, 4 tables, 1 algorithm)

This paper contains 12 sections, 8 equations, 6 figures, 4 tables, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. MOTIVATION
  3. METHOD AND FRAMEWORK
  4. Outline of Optimization Framework
  5. Formulation and Algorithm
  6. Theoretical Analysis of Model Update Optimization
  7. QUANTITATIVE EXPERIMENTS
  8. Datasets, Testbed and evaluation framework
  9. Analysis of Optimization Potential
  10. Effective Trade-offs by Multi-objective Heuristic Algorithm
  11. Evaluation of Model Update Optimization
  12. CONCLUSIONS

Figures (6)

  • Figure 1: Comparison between traditional methods and proposed MO-SAE.
  • Figure 2: Progressive clipping test.
  • Figure 7: Power consumption-Storage space Relationship.
  • Figure 10: Runtime-Power consumption Relationship.
  • Figure 13: Variation of F1 score with Increasing bin density.
  • ...and 1 more figures