Research map / Graph learning / GNN architectures
Subgraph and distance-aware GNNs
Boost GNN expressivity with identity, distance encoding, and subgraph information.
46 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2019–2026) · an approach in GNN architectures, keyword cluster “wl · expressive · power”
State History Benchmarks Tried Baselines Checklist Preprints Who Export
Where it stands
Quiet None of the tracked benchmarks has a report from 2025–26.
14 papers in 2023–24 and 12 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.0% to 2.4%.
No longer reported by papers here: PROTEINS (accuracy, last in 2024); NCI1 (accuracy, last in 2024); MUTAG (accuracy, last in 2024); IMDB-B (accuracy, last in 2024); D&D (accuracy, last in 2024); IMDB-M (accuracy, last in 2024).
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
- 2019 · Weisfeiler and Leman Go Neural: Higher-Order Graph Neural Networks (AAAI 2019 · 1,234 citations)
- 2020 · Distance Encoding: Design Provably More Powerful Neural Networks for Graph Representation Learning (NeurIPS 2020 · 114 citations)
- 2021 · Identity-aware Graph Neural Networks (AAAI 2021 · 140 citations)
- 2022 · KerGNNs: Interpretable Graph Neural Networks with Graph Kernels (AAAI 2022 · 77 citations)
- 2023 · Graph Neural Networks for Link Prediction with Subgraph Sketching (ICLR 2023 · 17 citations)
- 2024 · Union Subgraph Neural Networks (AAAI 2024 · 11 citations)
- 2025 · Expressive Power of Graph Neural Networks for (Mixed-Integer) Quadratic Programs (ICML 2025 · 0 citations)
- 2026 · The Logical Expressiveness of Topological Neural Networks (ICLR 2026 · 0 citations)
Earliest papers here
- Weisfeiler and Leman Go Neural: Higher-Order Graph Neural Networks (AAAI 2019 · 1,234 citations)
- How Powerful are Graph Neural Networks? (ICLR 2019 · 374 citations)
- Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels (NeurIPS 2019 · 126 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.
PROTEINS · accuracy (higher is better)
12 papers, 2019–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 77.2 | Provably Powerful Graph Networks (NeurIPS 2019 · 18 citations) |
| 2024 | 78.8 | Weisfeiler and Lehman Go Paths: Learning Topological Features via Path Complexes (AAAI 2024 · 7 citations) |
NCI1 · accuracy (higher is better)
11 papers, 2019–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 84.2 | Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels (NeurIPS 2019 · 126 citations) |
| 2024 | 85.1 | Weisfeiler and Lehman Go Paths: Learning Topological Features via Path Complexes (AAAI 2024 · 7 citations) |
MUTAG · accuracy (higher is better)
8 papers, 2019–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 90.55 | Provably Powerful Graph Networks (NeurIPS 2019 · 18 citations) |
| 2021 | 92.7 | Weisfeiler and Lehman Go Cellular: CW Networks (NeurIPS 2021 · 5 citations) |
IMDB-B · accuracy (higher is better)
6 papers, 2019–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 76.9 | Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels (NeurIPS 2019 · 126 citations) |
| 2024 | 79.6 | Exploring Consistency in Graph Representations: from Graph Kernels to Graph Neural Networks (NeurIPS 2024 · 1 citation) |
D&D · accuracy (higher is better)
5 papers, 2021–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 77.8 | Nested Graph Neural Networks (NeurIPS 2021 · 32 citations) |
| 2023 | 78.01 | Local Vertex Colouring Graph Neural Networks (ICML 2023 · 0 citations) |
IMDB-M · accuracy (higher is better)
5 papers, 2019–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 52.8 | Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels (NeurIPS 2019 · 126 citations) |
| 2024 | 56.53 | Exploring Consistency in Graph Representations: from Graph Kernels to Graph Neural Networks (NeurIPS 2024 · 1 citation) |
What has been tried
35 papers here report 348 measured changes (a component added, removed or swapped, compared with a baseline): helped 59%, no clear effect 7%, hurt 28%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| use union subgraphs instead of overlap subgraphs | MUTAG · accuracy | 85.70 ± 7.40 -> 87.31 ± 5.29 | Helped | Union Subgraph Neural Networks (AAAI 2024 · 11 citations) |
| Increase maximum lifting dimension from 2 to 7 | ZINC · test MAE | 0.160 ± 0.009 -> 0.096 ± 0.006 | Helped | Weisfeiler and Lehman Go Paths: Learning Topological Features via Path Complexes (AAAI 2024 · 7 citations) |
| add positional encodings | peptides-func · AP | 0.7275±0.0066 -> 0.7311\boldsymbol{±}0.0066 | Helped | Spatio-Spectral Graph Neural Networks (NeurIPS 2024 · 2 citations) |
| add equivariant eigenvector update to EPNN | MNIST Superpixel (k=3) · accuracy | 48.45 ± 1.2 % -> 60.95 ± 0.9 % | Helped | Spectral Graph Neural Networks are Incomplete on Graphs with a Simple Spectrum (NeurIPS 2025 · 0 citations) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
| Baseline | Benchmark | Metric | Reported | Papers |
|---|---|---|---|---|
| GraphSAGE | D&D | accuracy | 71.6 – 76.44 | 2 papers, e.g. NeurIPS 2021 and NeurIPS 2024 |
What reviewers will expect
Datasets and baselines named in the 8 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- GCN · 2 of 8 papers
- GAT · 2 of 8 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 26 papers; from OpenAlex and the papers' affiliations).
- Rheinisch Westfälische Technische Hochschule Aachen · 4 papers
- Peking University · 3 papers
- RWTH Aachen University · 2 papers
- University of Antwerp · 2 papers
- Massachusetts Institute of Technology · 2 papers
- University of Oxford · 2 papers
- Technion - Israel Institute of Technology · 2 papers
- Washington University · 2 papers
Use it
Download BibTeX of 35 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
- Weisfeiler-Leman expressivity hierarchies (66 papers)
- Substructure-aware GNN expressivity (50 papers)
Elsewhere in the topic
- Graph structure learning (201 papers)
- GNN generalization theory (187 papers)
- Simple and scalable GNNs (131 papers)
- Expressive equivariant message passing (109 papers)
- Robustness, privacy, and fairness of GNNs (94 papers)
- Theory of graph convolutional networks (83 papers)
- Deep graph convolution and simplification (82 papers)
- Physics-inspired GNNs (81 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).