Anchor-guided Hypergraph Condensation with Dual-level Discrimination
2026-05-11 • Machine Learning
Machine Learning
AI summaryⓘ
The authors address the problem of simplifying large hypergraphs to make training hypergraph neural networks easier and faster. They note that earlier methods didn't optimize the structure and features together, which led to less useful simplified hypergraphs. To fix this, the authors propose AHGCDD, which uses smart node initialization, an anchor-based way to create hyperedges, and a special learning objective to keep the important utility without extra heavy training. Their experiments show that this approach is both more effective and efficient than previous methods.
HypergraphHypergraph Neural Network (HNN)Hypergraph Condensation (HGC)Heat Kernel PageRank (HKPR)HyperedgeAnchor-guided synthesisDual-level discriminationFeature condensationGraph condensationTrajectory-based optimization
Authors
Fan Li, Xiaoyang Wang, Chen Chen, Wenjie Zhang
Abstract
The increasing prevalence of large-scale hypergraphs poses significant computational challenges for hypergraph neural network (HNN) training. To address this, hypergraph condensation (HGC) distills large real hypergraphs into compact yet informative synthetic ones, beyond graph condensation (GC) methods limited to pairwise relations. However, existing HGC methods rely on decoupled training architectures, where structure generators are pre-trained on the original hypergraph but not jointly optimized with condensed features during refinement, resulting in misaligned structures that degrade downstream utility. Moreover, trajectory-based optimization incurs substantial computational overhead in refinement, limiting condensation efficiency. To tackle these issues, we propose \textbf{A}nchor-guided \textbf{H}yper\textbf{G}raph \textbf{C}ondensation with \textbf{D}ual-level \textbf{D}iscrimination (\textbf{AHGCDD}), which consists of three key components: (1) a node initialization module based on Heat Kernel PageRank (HKPR) to encode structural knowledge into feature semantics; (2) an anchor-guided hyperedge synthesis strategy for joint optimization of condensed features and structure; (3) a theoretically grounded dual-level discrimination objective for utility-preserving condensation without redundant HNN training. Extensive experiments demonstrate the superior effectiveness and efficiency of AHGCDD.