Research map / Graph learning / Graph algorithms and clustering
Neural and local graph clustering
Cluster graphs with neural modularity, flow diffusion, and correlation clustering.
81 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Graph algorithms and clustering, keyword cluster “graph clustering · spectral clustering · clusters”
State History Benchmarks Tried Checklist Preprints Who Export
Where it stands
Still moving The best reported result is still improving on 2 of 4 tracked benchmarks.
21 papers in 2023–24 and 21 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 10.7% to 10.4%.
Open questions the data points to
- Cora (ACC): best went from 78.1 (up to 2024) to 79.19 in 2025 — One Node One Model: Featuring the Missing-Half for Graph Clustering (AAAI 2025 · 9 citations)
- Pubmed (NMI): best went from 32.9 (up to 2024) to 34.57 in 2025 — One Node One Model: Featuring the Missing-Half for Graph Clustering (AAAI 2025 · 9 citations)
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Community Detection on Evolving Graphs (NeurIPS 2016 · 18 citations)
- 2017 · Learning A Structured Optimal Bipartite Graph for Co-Clustering (NeurIPS 2017 · 86 citations)
- 2018 · Bayesian Robust Attributed Graph Clustering: Joint Learning of Partial Anomalies and Group Structure (AAAI 2018 · 76 citations)
- 2019 · Linkage Based Face Clustering via Graph Convolution Network (CVPR 2019 · 201 citations)
- 2020 · Adaptive Graph Convolutional Network With Attention Graph Clustering for Co-Saliency Detection (CVPR 2020 · 111 citations)
- 2021 · Multiple Kernel Clustering with Kernel k-Means Coupled Graph Tensor Learning (AAAI 2021 · 76 citations)
- 2022 · RepBin: Constraint-Based Graph Representation Learning for Metagenomic Binning (AAAI 2022 · 13 citations)
- 2023 · Hard Sample Aware Network for Contrastive Deep Graph Clustering (AAAI 2023 · 128 citations)
- 2024 · Attribute-Missing Graph Clustering Network (AAAI 2024 · 58 citations)
- 2025 · One Node One Model: Featuring the Missing-Half for Graph Clustering (AAAI 2025 · 9 citations)
- 2026 · Bipartite Graph Attention-based Clustering for Large-scale scRNA-seq Data (ICML 2026 · 0 citations)
Earliest papers here
- Community Detection on Evolving Graphs (NeurIPS 2016 · 18 citations)
- ReD-SFA: Relation Discovery Based Slow Feature Analysis for Trajectory Clustering (CVPR 2016 · 12 citations)
- Graph Clustering: Block-models and model free results (NeurIPS 2016 · 5 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
Cora · NMI (higher is better)
6 papers, 2023–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2023 | 62.28 | Dink-Net: Neural Clustering on Large Graphs (ICML 2023 · 6 citations) |
| 2024 | 63.97 | LSEnet: Lorentz Structural Entropy Neural Network for Deep Graph Clustering (ICML 2024 · 1 citation) |
Citeseer · NMI (higher is better)
5 papers, 2023–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2023 | 45.87 | Dink-Net: Neural Clustering on Large Graphs (ICML 2023 · 6 citations) |
| 2024 | 52.26 | LSEnet: Lorentz Structural Entropy Neural Network for Deep Graph Clustering (ICML 2024 · 1 citation) |
Cora · ACC (higher is better)
5 papers, 2023–2025 · still improving
| Year | Best so far | Set by |
|---|---|---|
| 2023 | 78.1 | Dink-Net: Neural Clustering on Large Graphs (ICML 2023 · 6 citations) |
| 2025 | 79.19 | One Node One Model: Featuring the Missing-Half for Graph Clustering (AAAI 2025 · 9 citations) |
Pubmed · NMI (higher is better)
4 papers, 2023–2025 · still improving
| Year | Best so far | Set by |
|---|---|---|
| 2023 | 32.9 | Scalable Attributed-Graph Subspace Clustering (AAAI 2023 · 20 citations) |
| 2025 | 34.57 | One Node One Model: Featuring the Missing-Half for Graph Clustering (AAAI 2025 · 9 citations) |
What has been tried
64 papers here report 650 measured changes (a component added, removed or swapped, compared with a baseline): helped 67%, no clear effect 10%, hurt 19%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add the dual non-contrastive clustering loss | Cora · ACC | 57.17 -> 66.65 | Helped | Attribute-Missing Graph Clustering Network (AAAI 2024 · 58 citations) |
| Use auxiliary-information loss with λ increased from 0.0 to 0.8 | Cora · NMI | 49.9 -> 62.1 | Helped | DGCLUSTER: A Neural Framework for Attributed Graph Clustering via Modularity Maximization (AAAI 2024 · 32 citations) |
| Vary the anchor rate | HandWritten4 and Mnist4 · Clustering metrics | Mixed | Tensorized Label Learning on Anchor Graph (AAAI 2024 · 18 citations) | |
| add feature decorrelation loss L_FD | ACM, DBLP, Amazon, AMiner · ACC | 0.9121 -> 0.9322; 0.8212 -> 0.8309; 0.5885 -> 0.6603; 0.6817 -> 0.7308 | Helped | Upper Bounding Barlow Twins: A Novel Filter for Multi-Relational Clustering (AAAI 2024 · 16 citations) |
| add feature-cross augmentation instead of omitting augmentation | Cora · ACC | 75.95 -> 79.19 | Helped | One Node One Model: Featuring the Missing-Half for Graph Clustering (AAAI 2025 · 9 citations) |
| add TopoReg to the MFC backbone | Enron, Highschool, DBLP, and Cora · ACC, NMI, ARI, and Modularity | Mixed | Learning Persistent Community Structures in Dynamic Networks via Topological Data Analysis (AAAI 2024 · 9 citations) | |
| Add higher-order neighbors to the order-1 neighbor matrix (orders 1–3) | COIL20 and Semeion · clustering metrics (ACC, NMI, PUR) | Helped | Fine-Grained Bipartite Concept Factorization for Clustering (CVPR 2024 · 9 citations) | |
| Use SAGE architecture rather than other tested architectures | Synthetic LFR networks · AMI | Helped | The Map Equation Goes Neural: Mapping Network Flows with Graph Neural Networks (NeurIPS 2024 · 3 citations) | |
| Vary the unigram sampling parameter across -1, 0, 0.25, 0.5, 0.75, and 1 | DC-SBM simulations · proportion of nodes correctly recovered | No clear effect | Community Detection Guarantees using Embeddings Learned by Node2Vec (NeurIPS 2024 · 2 citations) | |
| Replace the structural PPR kernel with an inner-product feature kernel | Synthetic structural covariate shifts · AE | Hurt | Adjusted Count Quantification Learning on Graphs (NeurIPS 2025 · 1 citation) | |
| Increase partitioning-tree height H | Cora, Citeseer, AMAP, and Computer · clustering performance | Helped | LSEnet: Lorentz Structural Entropy Neural Network for Deep Graph Clustering (ICML 2024 · 1 citation) | |
| add neural spectral clustering instead of numerical spectral clustering | BetaSeg test · DSC | 0.721 -> 0.816 | Helped | SCCS: Deep Neural Spectral Clustering for Self-Supervised Subcellular Structure Segmentation (AAAI 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 13 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- CiteSeer · 4 of 13 papers
- Cora · 3 of 13 papers
- Pubmed · 2 of 13 papers
- Twitch-Gamers · 2 of 13 papers
- mnist · 2 of 13 papers
- cifar10 · 2 of 13 papers
Baselines
- MVGRL · 2 of 13 papers
- CCGC · 2 of 13 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 42 papers; from OpenAlex and the papers' affiliations).
- National University of Defense Technology · 5 papers
- University of Illinois Chicago · 3 papers
- University of Electronic Science and Technology of China · 3 papers
- University of Waterloo · 2 papers
- National University of Singapore · 2 papers
Use it
Download BibTeX of 37 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Anchor-based multi-view clustering (31 papers)
Elsewhere in the topic
- Neural combinatorial optimization on graphs (159 papers)
- Graph comparison and random features (155 papers)
- Graph embedding and spectral sparsification (126 papers)
- Continuous DAG structure learning (102 papers)
- Graph distance and retrieval (79 papers)
- Deep learning for graph matching (42 papers)
- Optimization algorithms for graph matching (42 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).