Metadata relationship mining and intelligent bloodline tracing system empowered by graph neural networks

Metadata serves a vital function in assuring data transparency, traceability, and accountability throughout digital ecosystems. With the ever-growing amount of data and metadata produced, intelligent systems are increasingly demanded to mine the subtle latent relations of metadata entities and accurately trace lineage over time. Existing rule-based solutions and early graph- based models ignore the temporal semantic, heterogeneity, and hierarchical structure inherent to metadata environments. Recent advances in GNNs have been promising, but current implementations face difficulties incorporating heterogeneous or temporal data. In this research an innovative intelligent system, Temporal-Heterogeneous Hierarchical Graph Neural Network for Metadata Mining and Bloodline Tracing (TH-HGNN-MMBT) that analyse temporal, heterogeneous, hierarchical metadata data to achieve lineage traceability and discover relationships of metadata have been proposed. This model builds a hierarchical graph in which every entity such as dataset, file, user, step in metadata environment is taken as node, and their dependences such as ownership, latency, association is represented as time-aware heterogenous edges. The proposed TH-HGNN-MMBT model achieves a 12.8% gain in temporal consistency and a 7.5% gain in bloodline completeness filtering, processes metadata 34.2% faster, and has 6.8% better scaling for large-scale lineage analysis. The architecture enables discovering hidden relationships, inferring missing links, and rebuilding heterogenous metadata bloodline between layers of abstraction.

Authors

Publication Details

Journal
Scientific Reports
Published
2026-11-18
DOI
https://doi.org/10.1038/s41598-026-78455-9
Primary Topic
Advanced Graph Neural Networks
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Metadata relationship mining and intelligent bloodline tracing system empowered by graph neural networks

Jiaxing Xu, Xiaofei Gao, Fenghua Lv, Lei Lv
1 links
Scientific Reports
Advanced Graph Neural Networks
article

Metadata relationship mining and intelligent bloodline tracing system empowered by graph neural networks

Jiaxing Xu, Xiaofei Gao, Fenghua Lv, Lei Lv
article en

Abstract

Metadata serves a vital function in assuring data transparency, traceability, and accountability throughout digital ecosystems. With the ever-growing amount of data and metadata produced, intelligent systems are increasingly demanded to mine the subtle latent relations of metadata entities and accurately trace lineage over time. Existing rule-based solutions and early graph- based models ignore the temporal semantic, heterogeneity, and hierarchical structure inherent to metadata environments. Recent advances in GNNs have been promising, but current implementations face difficulties incorporating heterogeneous or temporal data. In this research an innovative intelligent system, Temporal-Heterogeneous Hierarchical Graph Neural Network for Metadata Mining and Bloodline Tracing (TH-HGNN-MMBT) that analyse temporal, heterogeneous, hierarchical metadata data to achieve lineage traceability and discover relationships of metadata have been proposed. This model builds a hierarchical graph in which every entity such as dataset, file, user, step in metadata environment is taken as node, and their dependences such as ownership, latency, association is represented as time-aware heterogenous edges. The proposed TH-HGNN-MMBT model achieves a 12.8% gain in temporal consistency and a 7.5% gain in bloodline completeness filtering, processes metadata 34.2% faster, and has 6.8% better scaling for large-scale lineage analysis. The architecture enables discovering hidden relationships, inferring missing links, and rebuilding heterogenous metadata bloodline between layers of abstraction.

Scientific Reports
Advanced Graph Neural Networks
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon