AdaE: Knowledge Graph Embedding With Adaptive Embedding Sizes

doi:10.1109/TKDE.2025.3566270

	AdaE: Knowledge Graph Embedding With Adaptive Embedding Sizes
	Guan, Zhanpeng 1,2; Zhang, Fuwei 3; Zhang, Zhao 1; Zhuang, Fuzhen 3,4; Wang, Fei 1; An, Zhulin 1; Xu, Yongjun 1
	2025-08-01
发表期刊	IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
ISSN	1041-4347
卷号	37 期号:8 页码:4432-4445
摘要	Knowledge Graph Embedding (KGE) aims to learn dense embeddings as the representations for entities and relations in KGs. Indeed, the entities in existing KGs suffer from the data imbalance issue, i.e., there exists a substantial disparity in the occurrence frequencies among various entities. Existing KGE models pre-define a unified and fixed dimension size for all entity embeddings. However, embedding sizes of entities are highly desired for their frequencies, while a uniform embedding size may result in inadequate expression of entities, i.e., leading to overfitting for low-frequency entities and underfitting for high-frequency ones. A straight-forward idea is to set the embedding sizes for each entity before KGE training. However, manually selecting different embedding sizes is labor-intensive and time-consuming, posing challenges in real-world applications. To tackle this problem, we propose AdaE, which adaptively learns KG embeddings with different embedding sizes during training. In particular, AdaE is capable of selecting appropriate dimension sizes for each entity from a continuous integer space. To this end, we specially tailor bilevel optimization for the KGE task, which alternately learns representations and embedding sizes of entities. Our framework is general and flexible, fitting various existing KGE models. Extensive experiments demonstrate the effectiveness and compatibility of AdaE.
关键词	Knowledge graphs Adaptation models Training Data models Search problems Vectors Overfitting Tail Optimization Tensors Knowledge graph embedding (KGE) Data imbalance issue Dimension selection
DOI	10.1109/TKDE.2025.3566270
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Program of China[2024YFF0729003] ; National Natural Science Foundation of China[62206266] ; National Natural Science Foundation of China[62176014] ; Fundamental Research Funds for the Central Universities
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Computer Science, Information Systems ; Engineering, Electrical & Electronic
WOS记录号	WOS:001525525600006
出版者	IEEE COMPUTER SOC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/42039
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Zhang, Zhao
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 3.Beihang Univ, Inst Artificial Intelligence, Beijing 100191, Peoples R China 4.Zhongguancun Lab, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Guan, Zhanpeng,Zhang, Fuwei,Zhang, Zhao,et al. AdaE: Knowledge Graph Embedding With Adaptive Embedding Sizes[J]. IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING,2025,37(8):4432-4445.
APA	Guan, Zhanpeng.,Zhang, Fuwei.,Zhang, Zhao.,Zhuang, Fuzhen.,Wang, Fei.,...&Xu, Yongjun.(2025).AdaE: Knowledge Graph Embedding With Adaptive Embedding Sizes.IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING,37(8),4432-4445.
MLA	Guan, Zhanpeng,et al."AdaE: Knowledge Graph Embedding With Adaptive Embedding Sizes".IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 37.8(2025):4432-4445.