Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Optimizing Value of Learning in Task-Oriented Federated Meta-Learning Systems

Bibo Wu, Fang Fang, Xianbin Wang

TL;DR

The paper tackles task personalization in federated learning by embedding meta-learning in a NOMA-based edge network. It introduces VoL to quantify per-device learning needs and TLW to encode task-specific importance and fairness, formulating a TLW-based VoL maximization as a non-convex MINLP. A PDQN-based solution handles the hybrid discrete-continuous scheduling and resource allocation, using an MDP with carefully designed state, action, and reward structures. Simulations on non-IID CIFAR-10 show the proposed method outperforms OMA and simple benchmarks, improving convergence, test accuracy, and overall VoL. The approach provides a scalable, privacy-preserving framework for task-oriented edge learning with efficient communication.

Abstract

Federated Learning (FL) has gained significant attention in recent years due to its distributed nature and privacy preserving benefits. However, a key limitation of conventional FL is that it learns and distributes a common global model to all participants, which fails to provide customized solutions for diverse task requirements. Federated meta-learning (FML) offers a promising solution to this issue by enabling devices to finetune local models after receiving a shared meta-model from the server. In this paper, we propose a task-oriented FML framework over non-orthogonal multiple access (NOMA) networks. A novel metric, termed value of learning (VoL), is introduced to assess the individual training needs across devices. Moreover, a task-level weight (TLW) metric is defined based on task requirements and fairness considerations, guiding the prioritization of edge devices during FML training. The formulated problem, to maximize the sum of TLW-based VoL across devices, forms a non-convex mixed-integer non-linear programming (MINLP) challenge, addressed here using a parameterized deep Q-network (PDQN) algorithm to handle both discrete and continuous variables. Simulation results demonstrate that our approach significantly outperforms baseline schemes, underscoring the advantages of the proposed framework.

Optimizing Value of Learning in Task-Oriented Federated Meta-Learning Systems

TL;DR

The paper tackles task personalization in federated learning by embedding meta-learning in a NOMA-based edge network. It introduces VoL to quantify per-device learning needs and TLW to encode task-specific importance and fairness, formulating a TLW-based VoL maximization as a non-convex MINLP. A PDQN-based solution handles the hybrid discrete-continuous scheduling and resource allocation, using an MDP with carefully designed state, action, and reward structures. Simulations on non-IID CIFAR-10 show the proposed method outperforms OMA and simple benchmarks, improving convergence, test accuracy, and overall VoL. The approach provides a scalable, privacy-preserving framework for task-oriented edge learning with efficient communication.

Abstract

Federated Learning (FL) has gained significant attention in recent years due to its distributed nature and privacy preserving benefits. However, a key limitation of conventional FL is that it learns and distributes a common global model to all participants, which fails to provide customized solutions for diverse task requirements. Federated meta-learning (FML) offers a promising solution to this issue by enabling devices to finetune local models after receiving a shared meta-model from the server. In this paper, we propose a task-oriented FML framework over non-orthogonal multiple access (NOMA) networks. A novel metric, termed value of learning (VoL), is introduced to assess the individual training needs across devices. Moreover, a task-level weight (TLW) metric is defined based on task requirements and fairness considerations, guiding the prioritization of edge devices during FML training. The formulated problem, to maximize the sum of TLW-based VoL across devices, forms a non-convex mixed-integer non-linear programming (MINLP) challenge, addressed here using a parameterized deep Q-network (PDQN) algorithm to handle both discrete and continuous variables. Simulation results demonstrate that our approach significantly outperforms baseline schemes, underscoring the advantages of the proposed framework.
Paper Structure (15 sections, 27 equations, 5 figures, 1 table, 1 algorithm)

This paper contains 15 sections, 27 equations, 5 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Problem Formulation
  3. FML Training
  4. Computation and Communication Models
  5. Value of Learning
  6. Task Level Weight
  7. Problem Formulation
  8. PDQN-based Device Scheduling and Resource Allocation Design
  9. MDP Model
  10. State space
  11. Action space
  12. Reward
  13. PDQN-based Algorithm
  14. Simulation Results
  15. Conclusion

Figures (5)

  • Figure 1: Federated meta-learning system model.
  • Figure 2: Training framework of PDQN.
  • Figure 3: Convergence of PDQN and DDPG algorithms.
  • Figure 4: FML performance on non-IID CIFAR-10 dataset.
  • Figure 5: VoL performance versus episode.