notesum.ai

Published at November 4

Against Multifaceted Graph Heterogeneity via Asymmetric Federated Prompt Learning

cs.LG

cs.AI

cs.DC

cs.SI

Released Date: November 4, 2024

Authors: Zhuoning Guo¹, Ruiqian Han¹, Hao Liu¹

Aff.: ¹The Hong Kong University of Science and Technology (Guangzhou)

Arxiv: http://arxiv.org/abs/2411.02003v1

Refer to caption

		Cora		CiteSeer		DBLP		Photo		Physics
Federated Method	Prompt Method	ACC	F1	ACC	F1	ACC	F1	ACC	F1	ACC	F1
	GPF	85.02	84.37	84.60	84.18	81.03	80.70	76.85	76.46	88.29	88.49
	ProG	86.45	85.91	85.72	85.85	82.20	82.01	71.47	70.47	86.73	86.98
Local	SUPT	85.66	85.23	87.00	86.88	80.57	80.52	73.31	72.87	89.69	89.56
	GPF	85.20	84.76	85.36	85.22	81.87	81.88	80.73	80.52	89.75	88.83
	ProG	86.70	86.13	86.49	86.05	82.43	82.21	72.21	71.76	89.05	89.27
FedAvg	SUPT	85.74	85.23	87.17	84.05	83.07	82.52	81.70	81.44	89.79	88.60
	GPF	80.52	79.74	83.20	82.69	80.57	79.84	82.93	82.79	87.60	87.19
	ProG	84.04	83.79	84.39	83.87	81.86	81.55	72.45	72.46	89.01	88.57
FedProx	SUPT	81.81	80.59	83.47	83.82	80.79	80.80	80.26	80.64	89.21	88.21
	GPF	77.08	76.39	75.79	74.87	78.57	77.63	74.43	74.37	85.20	84.85
	ProG	73.90	73.25	74.80	74.57	76.99	76.25	72.36	71.72	85.60	85.27
SCAFFOLD	SUPT	75.38	70.47	74.91	73.73	75.80	75.71	75.65	75.86	85.53	84.49
FedGPL		88.34	88.07	87.60	87.72	83.93	83.27	86.75	86.03	90.91	90.61