Efficient knowledge graph to text powered by LLGM: linear latent graph model

doi:10.1007/s40747-025-01985-8

	Efficient knowledge graph to text powered by LLGM: linear latent graph model
	Zhao, Xiaokang 1; Zheng, Yao 2; Shan, Yubo 2; Li, Jingyuan 1; Zhang, Kun 3; Wang, Yuanzhuo 4
	2025-08-01
发表期刊	COMPLEX & INTELLIGENT SYSTEMS
ISSN	2199-4536
卷号	11 期号:8 页码:22
摘要	Knowledge graph to text generation is crucial for interpreting complex structured data, yet state-of-the-art transformer models face significant computational burdens, limiting their practical deployment. This paper introduces the Linear Latent Graph Model (LLGM), a novel architecture that significantly enhances efficiency in KG-to-text generation without compromising performance. LLGM's core innovations are three-fold: (1) a Multi-head Statistical Attention (MSA) mechanism that achieves linear O(N) complexity by replacing pairwise token interactions with efficient statistical approximations, drastically reducing the primary computational bottleneck; (2) a Graph Latent Self-Attention (GLSA) module that efficiently encodes explicit graph structures using dimension-reduced intermediate representations, preserving relational fidelity with fewer parameters; and (3) a Graph Periodicity Projector (GPP) that optimizes feed-forward networks by decomposing representations into periodic and non-periodic components, adeptly capturing both regular and unique graph patterns. Experiments on the WebNLG and EventNarrative datasets demonstrate LLGM's significant contributions: it achieves competitive text generation quality, evidenced by a mere 0.8% BLEU-4 gap to the top model and the highest CIDEr score (4.63) on WebNLG, while requiring 20-37% fewer parameters than leading models. LLGM offers a robust and scalable solution, effectively bridging the efficiency-effectiveness gap in KG-to-text generation and enabling broader application in resource-constrained environments.
关键词	Knowledge graph to text generation Linear attention Low rank compression Fourier network
DOI	10.1007/s40747-025-01985-8
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[62172393] ; Henan Province Key Research and Development Project[241111211900]
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:001510573200009
出版者	SPRINGER HEIDELBERG
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/42366
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Li, Jingyuan
作者单位	1.Beijing Technol & Business Univ, Sch Comp & Artificial Intelligence, Beijing 100048, Peoples R China 2.Zhengzhou Univ, Henan Inst Adv Technol, Zhengzhou, Peoples R China 3.Tencent Inc, Pattern Recognit Ctr, WeChat AI, Beijing, Peoples R China 4.Chinese Acad Sci, Inst Comp Technol, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Zhao, Xiaokang,Zheng, Yao,Shan, Yubo,et al. Efficient knowledge graph to text powered by LLGM: linear latent graph model[J]. COMPLEX & INTELLIGENT SYSTEMS,2025,11(8):22.
APA	Zhao, Xiaokang,Zheng, Yao,Shan, Yubo,Li, Jingyuan,Zhang, Kun,&Wang, Yuanzhuo.(2025).Efficient knowledge graph to text powered by LLGM: linear latent graph model.COMPLEX & INTELLIGENT SYSTEMS,11(8),22.
MLA	Zhao, Xiaokang,et al."Efficient knowledge graph to text powered by LLGM: linear latent graph model".COMPLEX & INTELLIGENT SYSTEMS 11.8(2025):22.