Multi-view collaborative representation of intents over IBN: A heterogeneous graph aggregation method
Author Identifier (ORCID)
Abstract
Intent-Based Networking (IBN) enables the automatic deployment, correction, and optimization of network infrastructure configurations based on business intentions, thereby forming a closed-loop, autonomous, and business-centric network system. The comprehensive capture and accurate understanding of business intents are prerequisites for IBN, which, however, are challenging due to the inherent complexity and the temporal-spatial inconsistency of business objectives across diverse stakeholders. To this end, this paper introduces a hypergraph theory to represent multi-perspective intents in network businesses from a holistic viewpoint and proposes a Heterogeneous Isolated Graph Information Aggregation (HIGIA) model, unifying the following five key aspects. To express intents, knowledge graphs are employed to represent intents of different entities. Then, to address the semantic, structural, and temporal-spatial differences among diverse businesses, we design a Hybrid Mapping and Sharing Graph Attention Network (HMS-GAT) to capture the commonalities and differences among stakeholders’ demands. Next, to explore the inconsistency in the expression of intents across views, we incorporate the Gale–Shapley algorithm to enhance entity alignment. To address the lack of prior knowledge, bootstrapping is introduced to generate pseudo-alignment seeds for entity alignment; Finally, to mitigate missing or redundant information caused by heterogeneous expectations, we employ graph information aggregation to associate, complete, and refine intent descriptions, thus constructing a holistic hypergraph for business intents. Furthermore, to evaluate the hypergraph’s aggregation quality over all involved stakeholders, we propose information aggregation metrics based on information coverage and refinement. Experimental results demonstrate that HIGIA effectively captures service intents in IBN and provides a novel approach for multi-view heterogeneous intent aggregation.
Document Type
Journal Article
Date of Publication
1-1-2026
Publication Title
IEEE Transactions on Cognitive Communications and Networking
Publisher
IEEE
School
School of Engineering
Funders
National Research and Development Programs of China (2025YFB3109801) / National Natural Science Foundation of China (62361011, U24A20246) / Guizhou Provincial Science and Technology Plan Project (DXGA[2025]003, DXGA[2025]011) / Guizhou Province Scientist Workstation (KXJZ[2025]005) / Guizhou Province Science and Technology Innovation Platform (JSZX(2025)020) / Shanghai Pujiang Programme (24PJD117)
Copyright
subscription content
Comments
Liu, Y., Zou, S., Liwang, M., Ni, W., Wang, X., & Yang, C. (2026). Multi-view collaborative representation of intents over IBN: A heterogeneous graph aggregation method. IEEE Transactions on Cognitive Communications and Networking, 12, 5848–5863. https://doi.org/10.1109/TCCN.2026.3657062