Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Emergency Computing: An Adaptive Collaborative Inference Method Based on Hierarchical Reinforcement Learning

Weiqi Fu, Lianming Xu, Xin Wu, Li Wang, Aiguo Fei

TL;DR

The paper tackles the challenge of delivering timely, reliable emergency response services when infrastructure is compromised. It introduces the E-SC3I framework, which combines emergency computing, caching, integrated sensing and communication, and intelligence empowerment to enable rapid deployment and robust services under dynamic conditions. Central to this is ACIM, a hierarchical reinforcement learning approach that jointly optimizes model partitioning and per-layer pruning to minimize latency while meeting accuracy constraints in edge-cloud collaboration. Experimental results on CIFAR-10 with pretrained models show substantial latency reductions (up to 32%) and adaptive behavior under varying edge-compute resources, highlighting the framework's potential for real-time, resource-constrained emergency operations.

Abstract

In achieving effective emergency response, the timely acquisition of environmental information, seamless command data transmission, and prompt decision-making are crucial. This necessitates the establishment of a resilient emergency communication dedicated network, capable of providing communication and sensing services even in the absence of basic infrastructure. In this paper, we propose an Emergency Network with Sensing, Communication, Computation, Caching, and Intelligence (E-SC3I). The framework incorporates mechanisms for emergency computing, caching, integrated communication and sensing, and intelligence empowerment. E-SC3I ensures rapid access to a large user base, reliable data transmission over unstable links, and dynamic network deployment in a changing environment. However, these advantages come at the cost of significant computation overhead. Therefore, we specifically concentrate on emergency computing and propose an adaptive collaborative inference method (ACIM) based on hierarchical reinforcement learning. Experimental results demonstrate our method's ability to achieve rapid inference of AI models with constrained computational and communication resources.

Emergency Computing: An Adaptive Collaborative Inference Method Based on Hierarchical Reinforcement Learning

TL;DR

The paper tackles the challenge of delivering timely, reliable emergency response services when infrastructure is compromised. It introduces the E-SC3I framework, which combines emergency computing, caching, integrated sensing and communication, and intelligence empowerment to enable rapid deployment and robust services under dynamic conditions. Central to this is ACIM, a hierarchical reinforcement learning approach that jointly optimizes model partitioning and per-layer pruning to minimize latency while meeting accuracy constraints in edge-cloud collaboration. Experimental results on CIFAR-10 with pretrained models show substantial latency reductions (up to 32%) and adaptive behavior under varying edge-compute resources, highlighting the framework's potential for real-time, resource-constrained emergency operations.

Abstract

In achieving effective emergency response, the timely acquisition of environmental information, seamless command data transmission, and prompt decision-making are crucial. This necessitates the establishment of a resilient emergency communication dedicated network, capable of providing communication and sensing services even in the absence of basic infrastructure. In this paper, we propose an Emergency Network with Sensing, Communication, Computation, Caching, and Intelligence (E-SC3I). The framework incorporates mechanisms for emergency computing, caching, integrated communication and sensing, and intelligence empowerment. E-SC3I ensures rapid access to a large user base, reliable data transmission over unstable links, and dynamic network deployment in a changing environment. However, these advantages come at the cost of significant computation overhead. Therefore, we specifically concentrate on emergency computing and propose an adaptive collaborative inference method (ACIM) based on hierarchical reinforcement learning. Experimental results demonstrate our method's ability to achieve rapid inference of AI models with constrained computational and communication resources.
Paper Structure (15 sections, 4 equations, 3 figures, 1 table)

This paper contains 15 sections, 4 equations, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. E-SC3I Framework
  3. Emergency Computing
  4. Emergency Caching
  5. Integrated Communication and Sensing
  6. Intelligence Empowerment
  7. An Adaptive Collaborative Inference Method
  8. State
  9. Action
  10. Reward
  11. Loss Function
  12. Experiment
  13. Experimental Setup
  14. Performance Analysis
  15. Conclusion

Figures (3)

  • Figure 1: An Emergency Network with Sensing, Communication, Computation, Caching, and Intelligence.
  • Figure 2: An Adaptive Collaborative Inference Method for Emergency Computing.
  • Figure 3: Pruning Strategies with Different $R_{comp}$ .