Research map / Learning theory and algorithms / Sample complexity
Nonparametric Modeling of Higher-Order Interactions via Hypergraphons
Krishnakumar Balasubramanian
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use hypergraphon estimator instead of C3MM when 70% of hyperedges observed | hyperedge prediction · GPS (70% observed) | AUC: 0.69 -> 0.73 | Improved |
| use hypergraphon estimator instead of C3MM when 70% of hyperedges observed | hyperedge prediction · MovieLens (70% observed) | AUC: 0.81 -> 0.85 | Improved |
| use hypergraphon estimator instead of C3MM at 90% observed | hyperedge prediction · GPS (90% observed) | AUC: 0.83 -> 0.82 | No clear effect |
| use hypergraphon estimator instead of C3MM at 90% observed | hyperedge prediction · MovieLens (90% observed) | AUC: 0.90 -> 0.90 | No clear effect |
| use hypergraphon estimator instead of CMM at 70% observed | hyperedge prediction · GPS (70% observed) | AUC: 0.64 -> 0.73 | Improved |
| use spectral instead of random initialization (ablation) | SLH estimator · dense simulation (rho=1) | normalized L2 reconstruction error | No clear effect |
| use spectral instead of random initialization (ablation) | SLH estimator · sparse simulation (rho=0.7) | normalized L2 reconstruction error | Improved |
| increase number of nodes n at fixed k (ablation) | SLH estimator · well-specified simulation | normalized L2 reconstruction error | Improved |
| use rank-1 hypergraphon f=uvw (case 1) rather than logistic case 2 | SLH estimator · well-specified simulation | normalized L2 reconstruction error | Improved |
| increase number of blocks k from 0.1n^3 to 0.9n^3 (ablation) | SLH estimator · simulation n=20 dense | normalized L2 reconstruction error | Mixed |
| increase p1,q1 (richer misspecified hypergraphon) | SLH estimator · misspecified simulation n=100 | normalized L2 reconstruction error: 0.2 -> 0.8 | Hurt |
| set p2 equal to q2 (constant hypergraphon) | SLH estimator · misspecified simulation n=100 | normalized L2 reconstruction error | Improved |
Closest papers by meaning
- How Particle System Theory Enhances Hypergraph Message Passing (NeurIPS 2025 · similarity 0.86)
- Hypergraph Generation via Structured Stochastic Diffusion (NeurIPS 2026 · similarity 0.85)
- Higher-Order Graphon Neural Networks: Approximation and Cut Distance (ICLR 2025 · similarity 0.84)
- Higher-Order Graph Convolutional Network with Flower-Petals Laplacians on Simplicial Complexes (AAAI 2024 · similarity 0.83)
- DiPhon: Diffusion on Graphons for Scalable Graph Generation (NeurIPS 2026 · similarity 0.83)
- Beyond Graphs: Can Large Language Models Comprehend Hypergraphs? (ICLR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
