Fed-GAME: Personalized Federated Learning with Graph Attention Mixture-of-Experts For Time-Series Forecasting

Yi Li; Han Liu; Mingfeng Fan; Guo Chen; Chaojie Li; Biplab Sikdar

cs.LGMar 2, 2026

Fed-GAME: Personalized Federated Learning with Graph Attention Mixture-of-Experts For Time-Series Forecasting

Authors: Yi Li, Han Liu, Mingfeng Fan, Guo Chen, Chaojie Li, Biplab Sikdar

Organizations: School of Automation, Central South University · Department of Electrical and Computer Engineering, National University of Singapore · Department of Mechanical Engineering, National University of Singapore · School of Electrical Engineering and Telecommunications, University of New South Wales ·

Abstract

Federated learning (FL) on graphs shows promise for distributed time-series forecasting. Yet, existing methods rely on static topologies and struggle with client heterogeneity. We propose Fed-GAME, a framework that models personalized aggregation as message passing over a learnable dynamic implicit graph. The core is a decoupled parameter difference-based update protocol, where clients transmit parameter differences between their fine-tuned private model and a shared global model. On the server, these differences are decomposed into two streams: (1) averaged difference used to updating the global model for consensus (2) the selective difference fed into a novel Graph Attention Mixture-of-Experts (GAME) aggregator for fine-grained personalization. In this aggregator, shared experts provide scoring signals while personalized gates adaptively weight selective updates to support personalized aggregation. Experiments on two real-world electric vehicle charging datasets demonstrate that Fed-GAME outperforms state-of-the-art personalized FL baselines.

Figures & tables

QS =

\begin{cases}\frac{1}{n}\sum^{n}_{i=1}(1-q)\|y^{t}_{i}-\hat{y}^{t}_{i,q}\|,&y^{t}_{i}<\hat{y}^{t}_{i,q}\\ \frac{1}{n}\sum^{n}_{i=1}q\|y^{t}_{i}-\hat{y}^{t}_{i,q}\|,&y^{t}_{i}\geq\hat{y}^{t}_{i,q}\end{cases}

ICP =

\frac{1}{n}\sum^{n}_{i=1}\begin{cases}1,&\hat{y}^{t}_{i,q}\leq y^{t}_{i}\leq\hat{y}^{t}_{i,q^{{}^{\prime}}}\\ 0,&\text{otherwise}\end{cases}

MIL =

\frac{1}{n}\sum^{n}_{i=1}\|\hat{y}^{t}_{i,q}-\hat{y}^{t}_{i,q^{{}^{\prime}}}\|

Table 1: Evaluation Metrics

	Method	Aggregation	Palo Alto			Shenzhen
		Method	QS	MIL	ICP	QS	MIL	ICP
Step 6	No_FL	—	1.5228	4.7110	0.7585	18.6698	39.8802	0.6988
	LSTM-FedAvg	Average	1.4846	4.5517	0.7390	10.4788	10.1150	0.7381
	FedProx	Average	1.5378	4.7564	0.6571	9.8572	6.3500	0.6587
	pFedMe	Average	1.7047	5.0307	0.6677	10.5375	9.5557	0.7176
	PAG-FedAvg	Average	1.5304	4.2462	0.7534	13.7907	13.0350	0.4305

Table 2: Performance comparison of different FL forecasting models on Palo Alto and Shenzhen datasets

	Method	Aggregation Method	QS	MIL	ICP
Palo Alto	No_FL	—	2.477	7.031	0.6726
	Fed-GraphSAGE	GraphSAGE	1.8878	5.6907	0.756
	Fed-GAT	GAT	1.9081	5.6808	0.7462
	Fed-GAME	GAME	1.627	5.0027	0.7569
Shenzhen	No_FL	—	19.1557	35.0342	0.631
	Fed-GraphSAGE	GraphSAGE	11.9333	19.5901	0.756

Table 3: QS, MIL and ICP with different Aggregation Mechanism on different dataset (Prediction Step=12)

Federated Learning

Federated learning trains shared models across data holders without requiring them to pool their raw data.

See all 943 papers

FlatLand: Personalized Graph Federated Learning via Tailored Lorentz Space
Aug 21, 2026Jiahong Liu, Ram Samarth B B, Xinyu Fu +4Federated Graph Learning Federated Learning
Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China
FedIA: Importance-Aware Aggregation for Domain-Robust Federated Graph Learning
Sep 17, 2025Zhanting Zhou, Zeyu Ma, Ziqiang Zheng +1Federated Learning Aggregation Domain Generalization
University of Electronic Science and Technology of China, No. 2006, Xiyuan Ave, West Hi-Tech Zone, Chengdu, Sichuan 611731, China
F2STNet: Fair and Federated Spectral-Temporal Modeling for Graph Forecasting
Aug 10, 2026Jiayi Zhang, Jinfeng Xu, Hewei Wang +7Federated Learning Aggregation Federated Graph Learning
University of Nottingham · The University of Hong Kong · Carnegie Mellon University +1

Aug 21, 2026J/K move · Enter open · S save

FlatLand: Personalized Graph Federated Learning via Tailored Lorentz Space

Jiahong Liu, Ram Samarth B B, Xinyu Fu, Menglin Yang+3

Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China

Federated Graph Learning Federated Learning Heterogeneous Federated Learning Graph Representation Learning

Open paper AskPDF arXiv

Federated learning enables privacy-preserving collaborative training, but highly heterogeneous client data remain challenging, especially in graph federated learning where clients possess structurally diverse graphs. Existing personalized federated learning (PFL) methods ignore the intrinsic geometric properties of diverse graph structures. We propose FlatLand, a novel personalized federated learning method that embeds different clients' data in tailored Lorentz space of hyperbolic geometry. Our key insight is that hyperbolic geometry naturally accommodates the intrinsic negative curvature prevalent in real-world graphs, while the time-like dimension in Lorentz space provides a principled way to encode client-specific heterogeneity. We develop a parameter decoupling strategy that separates heterogeneous information (captured in time-like parameters) from common knowledge (preserved in space-like parameters), enabling direct aggregation without requiring client similarity estimation and extra calculation modules. Empirical results on diverse federated graph learning tasks demonstrate that FlatLand achieves superior performance, particularly in low-dimensional settings.

Abstract

Federated Learning

Explore similar work

FlatLand: Personalized Graph Federated Learning via Tailored Lorentz Space

FedIA: Importance-Aware Aggregation for Domain-Robust Federated Graph Learning

F2STNet: Fair and Federated Spectral-Temporal Modeling for Graph Forecasting