- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
The paper proposes HSENet for attributed hypergraph clustering. It formulates hypergraph structural entropy for hypergraphs with edge-dependent vertex weights and derives a continuous Lovász-extension relaxation for learning soft hierarchical assignments. The framework jointly trains masked feature reconstruction, hyperedge filling, representation-based topology enhancement, and an entropy-guided clustering module. Experiments compare HSENet with ten baselines on eight datasets. HSENet achieves the best reported mean performance across all four metrics on five datasets.
Critical strengths
S1: The proposed objective explicitly evaluates how entire hyperedges are partitioned while accommodating edge-dependent vertex weights. Its Lovász relaxation agrees with the discrete splitting costs at binary assignments and enables gradient-based optimization. This provides a concrete mechanism for connecting representation learning to hypergraph partition quality.
S2: In the empirical evaluation, improvements across multiple datasets and metrics, together with component ablations and missing-attribute experiments, provide evidence that the proposed method is effective.
Critical weaknesses
W1: The experiments do not compare existing methods that explicitly optimize graph partitions. The paper discusses LSEnet, DeSE, ASIL, and SCISE but evaluates none of them. These methods could be considered for evaluation through a hypergraph-to-graph conversion. Although conversion can change the partitioning objective, that difference makes the comparison informative for assessing the benefit of directly modeling hyperedge splitting costs. The evaluation does not compare against methods that directly optimize structural objectives, such as MinCutPool (Bianchi et al., 2020) and DMoN (Tsitsulin et al., 2023). Evaluating these methods on a graph expansion of the same hypergraphs would help establish the benefit of the proposed hyperedge-based entropy objective.
Bianchi, F. M., Grattarola, D., and Alippi, C. (2020). “Spectral Clustering with Graph Neural Networks for Graph Pooling.” ICML.
Tsitsulin, A., Palowitch, J., Perozzi, B., and Müller, E. (2023). “Graph Clustering with Graph Neural Networks.” JMLR, 24(127), 1–21.
W2: The ablations do not sufficiently isolate the contribution of HSE. The k-means variant removes the hierarchical assigner, HSE loss, and orthogonality loss together, and retrains the encoder using only the generative objectives (Appendix E.1). Consequently, this comparison changes several components simultaneously. Improvements over this variant are uneven: Table 2 reports relatively small gains on Cora and higher NMI for k-means on 20News, although larger gains occur on AMiner. A controlled comparison retaining the encoder, enhanced topology, and assignment architecture while replacing HSE with another clustering objective would more clearly establish the value of the principal contribution. The current results support the effectiveness of the overall system more strongly than the specific advantage of its entropy objective.
Questions for authors
None.
7 comments
Kind Reminder: Rebuttal and Reviewer Feedback
Dear Reviewer,
Thank you for your valuable comments. As the discussion deadline is approaching, we would greatly appreciate it if you could revisit our rebuttal. We have added the required comparisons, in which HSENet and the HSE objective consistently achieve the highest performance. We hope these results address your concerns.
Please let us know if you have any remaining concerns. Thank you again for your time and consideration.
W1: Detailed comparison with graph partitioning methods
Thank you again for this helpful suggestion. Following our brief response, which referred to the updated manuscript PDF, we provide a more detailed response to W1 below, including the complete results added in Section 6.3 and Appendix E.1.
We agree that graph partitioning methods provide an informative comparison despite the change in objective induced by conversion. We evaluated all six requested methods: MinCutPool and DMoN, which optimize cut- and modularity-based objectives, and the graph structural entropy methods LSEnet, DeSE, ASIL, and SCISE. All graph baselines use binary clique expansions of the same eight hypergraphs; HSENet operates directly on hypergraphs.
Table R1. Complete results (%), mean±standard deviation over five runs. Bold indicates the highest mean in each row. F1 denotes Macro-F1; OOM denotes out of memory.
| Dataset | Metric | MinCutPool | DMoN | SCISE | DeSE | LSEnet | ASIL | HSENet |
|---|---|---|---|---|---|---|---|---|
| Cora | NMI | 27.8±14.0 | 26.8±13.8 | 53.1±1.1 | 39.8±3.7 | 29.4±7.3 | 32.0±3.8 | 57.5±0.8 |
| ARI | 19.8±10.2 | 19.8±10.6 | 48.9±1.1 | 34.4±5.0 | 21.7±9.4 | 24.4±3.3 | 52.9±0.9 | |
| ACC | 45.4±9.6 | 42.8±8.6 | 71.7±0.8 | 58.9±6.8 | 41.0±11.2 | 45.6±2.6 | 74.7±1.6 | |
| F1 | 40.7±17.8 | 29.2±12.5 | 70.1±0.9 | 46.6±10.8 | 21.3±13.6 | 27.3±3.3 | 69.5±4.0 | |
| Citeseer | NMI | 20.5±2.4 | 18.9±2.4 | 39.9±0.6 | 22.1±2.3 | 24.4±5.5 | 22.4±1.9 | 45.6±0.6 |
| ARI | 17.3±2.4 | 15.9±1.4 | 39.2±0.7 | 18.2±2.5 | 21.8±6.4 | 20.9±1.8 | 47.9±0.7 | |
| ACC | 44.4±2.4 | 41.4±2.4 | 64.1±0.9 | 47.7±4.0 | 46.7±7.8 | 45.9±3.5 | 72.1±0.4 | |
| F1 | 42.6±2.5 | 28.8±4.6 | 59.6±1.0 | 40.6±3.8 | 34.2±11.1 | 31.1±5.1 | 61.7±0.4 | |
| Pubmed | NMI | 20.6±1.1 | 0.2±0.2 | 33.6±0.5 | 17.3±8.3 | 25.1±7.5 | 24.4±2.5 | 35.2±2.0 |
| ARI | 20.8±2.3 | 0.2±0.3 | 38.4±1.0 | 15.8±7.9 | 25.5±9.8 | 22.9±3.6 | 40.1±3.0 | |
| ACC | 58.5±5.9 | 41.9±1.5 | 72.8±0.4 | 57.9±8.9 | 63.7±8.6 | 55.6±0.6 | 74.0±2.0 | |
| F1 | 57.8±5.4 | 24.5±6.4 | 70.9±0.3 | 41.8±8.5 | 51.6±11.1 | 45.4±1.6 | 72.1±1.8 | |
| DBLP | NMI | 53.2±2.4 | 38.7±8.4 | 59.4±2.9 | OOM | OOM | OOM | 65.8±2.6 |
| ARI | 45.6±3.2 | 31.3±8.4 | 56.3±4.5 | OOM | OOM | OOM | 61.8±5.1 | |
| ACC | 66.5±4.9 | 52.9±6.9 | 72.3±6.1 | OOM | OOM | OOM | 78.6±5.5 | |
| F1 | 60.8±13.2 | 37.7±8.4 | 69.9±6.5 | OOM | OOM | OOM | 78.5±4.9 | |
| 20News | NMI | OOM | OOM | 29.7±1.5 | OOM | OOM | OOM | 43.5±1.2 |
| ARI | OOM | OOM | 30.8±2.0 | OOM | OOM | OOM | 50.3±1.2 | |
| ACC | OOM | OOM | 61.6±1.2 | OOM | OOM | OOM | 76.0±1.0 | |
| F1 | OOM | OOM | 53.9±1.6 | OOM | OOM | OOM | 73.0±1.7 | |
| Mushroom | NMI | 4.5±9.0 | 0.0±0.0 | 21.4±0.4 | OOM | OOM | OOM | 45.8±0.1 |
| ARI | 5.9±11.8 | 0.0±0.0 | 26.8±0.4 | OOM | OOM | OOM | 54.2±0.1 | |
| ACC | 56.9±10.1 | 51.8±0.0 | 75.9±0.2 | OOM | OOM | OOM | 86.8±0.0 | |
| F1 | 42.7±17.2 | 34.1±0.0 | 75.8±0.2 | OOM | OOM | OOM | 86.7±0.0 | |
| AMiner | NMI | 33.0±1.0 | 38.8±5.2 | 42.9±0.3 | OOM | OOM | 45.9±4.9 | 49.9±0.9 |
| ARI | 17.1±0.8 | 27.6±3.6 | 28.1±0.7 | OOM | OOM | 31.2±4.6 | 35.4±0.5 | |
| ACC | 29.4±1.3 | 36.0±3.5 | 35.7±0.6 | OOM | OOM | 41.1±3.7 | 43.1±0.6 | |
| F1 | 27.4±1.3 | 23.7±1.0 | 27.3±0.5 | OOM | OOM | 17.7±2.7 | 22.1±1.4 | |
| Zoo | NMI | 0.0±0.0 | 0.0±0.0 | 72.8±1.9 | 0.0±0.0 | 14.1±0.6 | 10.9±0.7 | 90.5±1.9 |
| ARI | 0.0±0.0 | 0.0±0.0 | 64.1±4.1 | 0.0±0.0 | 4.1±0.3 | 2.3±0.5 | 91.8±3.2 | |
| ACC | 40.6±0.0 | 40.6±0.0 | 73.3±3.3 | 40.6±0.0 | 45.0±0.8 | 31.5±0.5 | 93.1±0.6 | |
| F1 | 8.3±0.0 | 8.3±0.0 | 63.0±2.3 | 8.3±0.0 | 22.4±2.4 | 18.9±0.2 | 78.7±1.8 |
HSENet achieves the highest mean NMI, ARI, and ACC among methods with available results on each of the eight datasets, and the highest Macro-F1 on six. The exceptions are Cora (SCISE: 70.1 vs. 69.5) and AMiner (MinCutPool: 27.4 and SCISE: 27.3 vs. 22.1). On Mushroom and 20News, its gains over the strongest available graph baseline are 24.4 and 13.8 percentage points in NMI, and 27.4 and 19.5 points in ARI, respectively.
Clique expansion yields dense graphs on Mushroom (21.1%) and 20News (20.2%). Zoo becomes a complete graph, whose cut costs cannot distinguish vertex groups of equal size; this loss of discriminative topology may contribute to the low NMI of several baselines. SCISE completes all eight datasets; MinCutPool and DMoN complete seven. These outcomes characterize the memory limitations of the graph clustering methods on the evaluated hypergraph datasets.
Overall, these results support HSENet's overall clustering advantage over the evaluated clique-expansion pipelines and are consistent with the benefit of directly modeling hypergraph structure. We hope these additions address the requested comparison with methods that explicitly optimize graph partitions.
W1: Comparison with graph partitioning methods
Thank you for this suggestion. The added experiments show that HSENet achieves the highest mean NMI and ACC among methods with available results on all 8 datasets.
We added a new subsection, “Structural Objectives and Graph Clustering Baselines,” where we evaluate all 6 requested methods: MinCutPool, DMoN, LSEnet, DeSE, ASIL, and SCISE. The graph baselines operate on binary clique expansions of the same hypergraphs, while HSENet operates directly on hypergraphs. This comparison includes both cut-/modularity-based methods and graph structural entropy methods.
Table R1. Graph clustering comparison. Results (%) are mean ± standard deviation over five runs.
HSENet achieves the highest NMI and ACC among methods with available results on each dataset, with NMI gains of 24.4 and 13.8 percentage points over the strongest available graph baseline on Mushroom and 20News, respectively. On these 2 datasets, expansion produces dense graphs (21.1% and 20.2% density), and DeSE, LSEnet, and ASIL encounter OOM on both. Zoo becomes a complete graph, whose cut costs depend only on cluster sizes and cannot distinguish equally sized vertex groups; this loss of discriminative topology may contribute to the low NMI of several baselines. These comparisons support HSENet's overall clustering advantage over clique-expansion pipelines. More results are added in Section 6.3 and Appendix E.2.
W2: Controlled evaluation of the HSE objective
The controlled objective replacements show that HSE achieves the highest average scores on all 4 metrics and the best average rank under the fixed HSENet training configurations.
We agree that the original k-means variant changes several components simultaneously and therefore evaluates the combined assignment-learning design. We have added the requested comparison: retaining the encoder, hierarchical assignment architecture, topology-enhancement procedure, reconstruction and hyperedge-filling losses, and orthogonality regularizer, we replace only the structural objective with differentiable hypergraph normalized cut (NeurIPS 07 & 17), strict hypergraph modularity (Science Advances 21), and Graph SE. For Graph SE, a weighted clique expansion of the enhanced EDVW hypergraph is used solely to evaluate the loss.
Table R2. Controlled structural-objective comparison. Scores (%) are equally averaged over 8 datasets.
HSE achieves the highest average scores across the 8 datasets on all 4 metrics, with particularly substantial gains over hypergraph modularity, ranging from 5.59 to 16.58 percentage points. HSE also achieves the best average rank of 1.56, whereas Graph SE and hypergraph NCut tie at 2.50. HSE's gains over Graph SE may reflect the benefits of directly modeling hyperedge splitting costs. Together, these results support HSE as an effective structural objective and indicate that it is an important contributor to HSENet's overall clustering performance. This comparison is added in Section 6.3. Objective definitions and detailed results are added in Appendix D.4 and E.1.