Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

UAV-assisted Unbiased Hierarchical Federated Learning: Performance and Convergence Analysis

Ruslan Zhagypar, Nour Kouzayha, Hesham ElSawy, Hayssam Dahrouj, Tareq Y. Al-Naffouri

TL;DR

The paper examines the theoretical convergence of the proposed unbiased UAV-assisted HFL algorithm under adverse channel conditions and highlights the impact of the limited battery capacity of the UAV on the efficiency of the HFL algorithm.

Abstract

The development of the sixth generation (6G) of wireless networks is bound to streamline the transition of computation and learning towards the edge of the network. Hierarchical federated learning (HFL) becomes, therefore, a key paradigm to distribute learning across edge devices to reach global intelligence. In HFL, each edge device trains a local model using its respective data and transmits the updated model parameters to an edge server for local aggregation. The edge server, then, transmits the locally aggregated parameters to a central server for global model aggregation. The unreliability of communication channels at the edge and backhaul links, however, remains a bottleneck in assessing the true benefit of HFL-empowered systems. To this end, this paper proposes an unbiased HFL algorithm for unmanned aerial vehicle (UAV)-assisted wireless networks that counteracts the impact of unreliable channels by adjusting the update weights during local and global aggregations at UAVs and terrestrial base stations (BS), respectively. To best characterize the unreliability of the channels involved in HFL, we adopt tools from stochastic geometry to determine the success probabilities of the local and global model parameter transmissions. Accounting for such metrics in the proposed HFL algorithm aims at removing the bias towards devices with better channel conditions in the context of the considered UAV-assisted network.. The paper further examines the theoretical convergence guarantee of the proposed unbiased UAV-assisted HFL algorithm under adverse channel conditions. One of the developed approach's additional benefits is that it allows for optimizing and designing the system parameters, e.g., the number of UAVs and their corresponding heights. The paper results particularly highlight the effectiveness of the proposed unbiased HFL scheme as compared to conventional FL and HFL algorithms.

UAV-assisted Unbiased Hierarchical Federated Learning: Performance and Convergence Analysis

TL;DR

The paper examines the theoretical convergence of the proposed unbiased UAV-assisted HFL algorithm under adverse channel conditions and highlights the impact of the limited battery capacity of the UAV on the efficiency of the HFL algorithm.

Abstract

The development of the sixth generation (6G) of wireless networks is bound to streamline the transition of computation and learning towards the edge of the network. Hierarchical federated learning (HFL) becomes, therefore, a key paradigm to distribute learning across edge devices to reach global intelligence. In HFL, each edge device trains a local model using its respective data and transmits the updated model parameters to an edge server for local aggregation. The edge server, then, transmits the locally aggregated parameters to a central server for global model aggregation. The unreliability of communication channels at the edge and backhaul links, however, remains a bottleneck in assessing the true benefit of HFL-empowered systems. To this end, this paper proposes an unbiased HFL algorithm for unmanned aerial vehicle (UAV)-assisted wireless networks that counteracts the impact of unreliable channels by adjusting the update weights during local and global aggregations at UAVs and terrestrial base stations (BS), respectively. To best characterize the unreliability of the channels involved in HFL, we adopt tools from stochastic geometry to determine the success probabilities of the local and global model parameter transmissions. Accounting for such metrics in the proposed HFL algorithm aims at removing the bias towards devices with better channel conditions in the context of the considered UAV-assisted network.. The paper further examines the theoretical convergence guarantee of the proposed unbiased UAV-assisted HFL algorithm under adverse channel conditions. One of the developed approach's additional benefits is that it allows for optimizing and designing the system parameters, e.g., the number of UAVs and their corresponding heights. The paper results particularly highlight the effectiveness of the proposed unbiased HFL scheme as compared to conventional FL and HFL algorithms.
Paper Structure (18 sections, 8 theorems, 93 equations, 13 figures, 1 table)

This paper contains 18 sections, 8 theorems, 93 equations, 13 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Contributions and Organization
  4. System Model
  5. Network model
  6. Channel Model
  7. Association Policy and SINR
  8. Proposed UAV-assisted Unbiased HFL
  9. Success Probabilities Derivation
  10. Performance analysis
  11. Proof of Convergence
  12. Latency Analysis
  13. Numerical results
  14. Conclusion
  15. Proof of Theorem 1
  16. ...and 3 more sections

Key Result

Theorem 1

In a UAV-assisted wireless network implementing HFL, the edge success probability for device $k, k\in S_u, u=1,2,...N_u$, describing the reliability of local model parameters transmission between the device and its serving UAV acting as edge server is defined as the probability that the SINR values where $l_k$ is the 3-D distance between a selected pair of device $k$ and a UAV. $\mathcal{P}_k^{z,

Figures (13)

  • Figure 1: System model representation with $N_u=4$ and $N_d=12$.
  • Figure 2: System model realization with $N_u=10$ and $N_d=50$.
  • Figure 3: Representation of HFL with a focus on the communication links.
  • Figure 4: Illustration of the improvement of upper bound as a function of $B_2$ and $B_3$. The blue plane represents attainable bound improvement, while the gray plane is for the reference and means no improvement.
  • Figure 5: Illustration of two pairs of devices and UAVs used to represent the success probabilities of the edge and backhaul links. The placements are taken from the system realization in Fig. \ref{['System_model_tex']}.
  • ...and 8 more figures

Theorems & Definitions (10)

  • Definition 1
  • Definition 2
  • Theorem 1: Edge success probability
  • Lemma 1
  • Lemma 2
  • Theorem 2: Backhaul success probability
  • Lemma 3
  • Theorem 3
  • Corollary 1
  • Lemma 4