Multi-Model Federated Learning Optimization Based on Multi-Agent Reinforcement Learning
Bibliographic record
Abstract
This paper addresses the problem of Multi-Model Federated Learning (MMFL) in a typical wireless network, where a cellular Base Station (BS) cooperates with multiple clients to simultaneously train several Machine Learning (ML) models. Accordingly, the objective of this paper is to make an efficient joint decision for client association and communication-computation resource allocation to optimize the performance of the MMFL algorithm. In this regard, an optimization problem is formulated to minimize the average global loss of ML models under clients' energy and delay constraints. It is shown that the problem is a mixed-integer optimization whose objective is implicit in terms of the decision variables. To solve the optimization problem, we propose a Multi-Agent Multi-Model Federated Learning (MAMMFL) scheme based on a cooperative multi-agent configuration to intelligently assign models and resources to clients. Specifically, the problem is first converted to a Markov Decision Process (MDP) problem, then it is divided into four sub-MDP problems, where each problem relates to a phase in MMFL. The reinforcement learning algorithm solves each sub-problem, and a team-Q algorithm is adopted to coordinate agents in a cooperative multi-agent setting. Simulation results show that the proposed method can outperform other baselines in terms of average global loss and resource consumption.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".