HyperFuse speeds up generating node embeddings for complex hypergraphs
HyperFuse: Fast Self-Supervised Node Embeddings for Attributed Hypergraphs
Machine Learning
Summary
Getting useful fingerprints (embeddings) for nodes in hypergraphs normally takes a long time because existing methods use complex models and many training cycles. The authors present HyperFuse, a new way to quickly create these fingerprints without needing labels or long training. HyperFuse uses a clever method to capture structure and node features efficiently, then trains a small model much faster. It works well across many real-world hypergraphs and is much faster than previous approaches while maintaining good accuracy.
What this means in practice
- •For data engineers: Rapidly generate node embeddings for many or changing hypergraphs to keep data models up to date without long retraining delays.
- •For graph analytics teams: Improve speed and accuracy of hypergraph-based classification and clustering tasks in domains like social networks, biology, or recommendation systems.
Authors
Megha P, Harshit Kumar, Srajan Agarwal, Anirban Banerjee, Olaf Wolkenhauer, Saptarshi Bej
Abstract
Self-supervised hypergraph representation learning can produce informative node embeddings, but existing methods often require deep encoders trained for hundreds of epochs, making embedding generation costly even for hypergraphs with a few thousand nodes. This limits applications requiring embeddings for many or evolving hypergraphs. We present HyperFuse, a label-free pipeline for fast hypergraph representation learning. HyperFuse (i) computes structural node coordinates by maximizing a spectral relaxation of hypergraph modularity using Banerjee's hypergraph adjacency and a matrix-free operator with cost linear in node-hyperedge incidences; (ii) constructs multi-scale feature summaries and assigns bounded utility weights to hyperedges based on member stability under feature and membership masking; and (iii) trains a lightweight utility-weighted hypergraph encoder for 100 epochs using an invariance-decorrelation objective. We compare HyperFuse with TriCL, SE-HSSL, VilLain, and HypeBoy on nine public hypergraphs using six downstream classifiers and k-means clustering. On the eight datasets where all methods completed, HyperFuse required 8.7 s per dataset on average, achieving 13-179x geometric-mean speed-ups over the baselines. It achieved the highest average accuracy with five of six classifiers, while classification and clustering performance was not significantly different from TriCL and SE-HSSL. Compared with HypeBoy, HyperFuse was 13x faster and 2.1-4.1 percentage points more accurate across all classifiers. HyperFuse provides a practical approach for fast, repeated hypergraph embedding generation.