Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Communication Efficient Cooperative Edge AI via Event-Triggered Computation Offloading

You Zhou, Changsheng You, Kaibin Huang

TL;DR

A channel-adaptive, event-triggered edge-inference framework that prioritizes efficient rare-event processing and effectively reduces communication overhead, as opposed to existing edge-inference approaches is proposed.

Abstract

Rare events, despite their infrequency, often carry critical information and require immediate attentions in mission-critical applications such as autonomous driving, healthcare, and industrial automation. The data-intensive nature of these tasks and their need for prompt responses, combined with designing edge AI (or edge inference), pose significant challenges in systems and techniques. Existing edge inference approaches often suffer from communication bottlenecks due to high-dimensional data transmission and fail to provide timely responses to rare events, limiting their effectiveness for mission-critical applications in the sixth-generation (6G) mobile networks. To overcome these challenges, we propose a channel-adaptive, event-triggered edge-inference framework that prioritizes efficient rare-event processing. Central to this framework is a dual-threshold, multi-exit architecture, which enables early local inference for rare events detected locally while offloading more complex rare events to edge servers for detailed classification. To further enhance the system's performance, we developed a channel-adaptive offloading policy paired with an online algorithm to dynamically determine the optimal confidence thresholds for controlling offloading decisions. The associated optimization problem is solved by reformulating the original non-convex function into an equivalent strongly convex one. Using deep neural network classifiers and real medical datasets, our experiments demonstrate that the proposed framework not only achieves superior rare-event classification accuracy, but also effectively reduces communication overhead, as opposed to existing edge-inference approaches.

Communication Efficient Cooperative Edge AI via Event-Triggered Computation Offloading

TL;DR

A channel-adaptive, event-triggered edge-inference framework that prioritizes efficient rare-event processing and effectively reduces communication overhead, as opposed to existing edge-inference approaches is proposed.

Abstract

Rare events, despite their infrequency, often carry critical information and require immediate attentions in mission-critical applications such as autonomous driving, healthcare, and industrial automation. The data-intensive nature of these tasks and their need for prompt responses, combined with designing edge AI (or edge inference), pose significant challenges in systems and techniques. Existing edge inference approaches often suffer from communication bottlenecks due to high-dimensional data transmission and fail to provide timely responses to rare events, limiting their effectiveness for mission-critical applications in the sixth-generation (6G) mobile networks. To overcome these challenges, we propose a channel-adaptive, event-triggered edge-inference framework that prioritizes efficient rare-event processing. Central to this framework is a dual-threshold, multi-exit architecture, which enables early local inference for rare events detected locally while offloading more complex rare events to edge servers for detailed classification. To further enhance the system's performance, we developed a channel-adaptive offloading policy paired with an online algorithm to dynamically determine the optimal confidence thresholds for controlling offloading decisions. The associated optimization problem is solved by reformulating the original non-convex function into an equivalent strongly convex one. Using deep neural network classifiers and real medical datasets, our experiments demonstrate that the proposed framework not only achieves superior rare-event classification accuracy, but also effectively reduces communication overhead, as opposed to existing edge-inference approaches.
Paper Structure (34 sections, 6 theorems, 53 equations, 7 figures, 1 algorithm)

This paper contains 34 sections, 6 theorems, 53 equations, 7 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. Models and Metrics
  3. Inference Model
  4. CNN Classification with Early Exiting
  5. Local Inference Energy Consumption
  6. Communication Model
  7. Performance Metrics
  8. Classification Accuracy
  9. Mobile Energy Consumption
  10. Overview of Dynamic Co-inference
  11. Model 1 - Local Inference
  12. Model 2 - Event-Triggered Computation Offloading
  13. Dual-threshold Based Event Detection - Scheme and Tradeoff
  14. Event Detection Scheme
  15. Head Event Detection
  16. ...and 19 more sections

Key Result

Lemma 1

(Feasibility Condition). Given the total number of events $M$, the bandwidth $B$, and the energy constraint $\xi$, the system is feasible to support offloading if the SNR satisfies the following condition:

Figures (7)

  • Figure 1: Edge co-inference system.
  • Figure 2: Illustration of a backbone model with early exiting.
  • Figure 3: Dual-threshold confidence aware inference progress.
  • Figure 4: Missing probability versus offloading constraints in the case of imbalanced ratio $R=4$.
  • Figure 5: Missing probability versus offloading constraints in the cases of imbalanced ratio $R=4$ and $R=9$.
  • ...and 2 more figures

Theorems & Definitions (8)

  • Definition 1
  • Lemma 1
  • Lemma 2
  • Lemma 3
  • Lemma 4
  • Proposition 1
  • Remark 1
  • Proposition 2