AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

Graph comparison and random features

Compare graphs with optimal transport and efficient graph random features.

155 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Graph algorithms and clustering, keyword cluster “search · directed · causal”

State History Tried Checklist Preprints Who Export

Where it stands

Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.

31 papers in 2023–24 and 41 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 15.8% to 20.4%.

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

2016: 9162017: 11172018: 14182019: 16192020: 14202021: 4212022: 15222023: 14232024: 17242025: 15252026: 2626

Most cited paper of each year

Earliest papers here

What has been tried

114 papers here report 906 measured changes (a component added, removed or swapped, compared with a baseline): helped 65%, no clear effect 10%, hurt 19%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.

Ablations in recent papers

What was changedMeasured onResultEffectPaper
Replace Full-STC with ESTCflr-l and trn1-l · path costHelpedLarge-Scale Multi-Robot Coverage Path Planning via Local Search (AAAI 2024 · 14 citations)
add progressive samplingOGBN-MAG · Test accuracy47.66 ± 0.45 -> 54.83±0.20HelpedLong-range Meta-path Search on Large-scale Heterogeneous Graphs (NeurIPS 2024 · 9 citations)
Increase α from 1/3 to 0.99Facebook · densityHelpedOptimal Quasi-clique: Hardness, Equivalence with Densest-k-Subgraph, and Quasi-partitioned Community Mining (AAAI 2024 · 5 citations)
Use GSG instead of RBF while retaining random samplingSIR, SW · localization error7.4±0.6 -> 6.6±1.5; -10.8%HelpedMultiple-Source Localization from a Single-Snapshot Observation Using Graph Bayesian Optimization (AAAI 2024 · 3 citations)
Disable the DOMV and MC reduction rulesh 141, h 165, h 181 · objective value gap over search timeHurtThreshold-Based Responsive Simulated Annealing for Directed Feedback Vertex Set Problem (AAAI 2024 · 3 citations)
Partition the space with L > 1 instead of L = 1SIFT10M, Tiny5M, and GIST · recall-QPS performanceHelpedProbabilistic Routing for Graph-Based Approximate Nearest Neighbor Search (ICML 2024 · 1 citation)
Use 2 or 4 partitions instead of the unpartitioned baselineSIFT1M · QPS-recall performanceHelpedCSPG: Crossing Sparse Proximity Graphs for Approximate Nearest Neighbor Search (NeurIPS 2024 · 1 citation)
add the relation netArcPie Seen · PL-F1 ssa0.693 -> 0.854Helped3DPGS: 3D Probabilistic Graph Search for Archaeological Piece Grouping (AAAI 2025 · 0 citations)
Use Φ1 or Φ2 instead of Φ0 in OSTDocument classification · classification performanceHelpedAn Efficient Orlicz-Sobolev Approach for Transporting Unbalanced Measures on a Graph (NeurIPS 2025 · 0 citations)
Add antithetic termination to repelling random walkscylinder · angular error (1-cos θ)0.0466(3) -> 0.0459(2)HelpedRepelling Random Walks (ICLR 2024 · 0 citations)
Train a neural modulation function instead of using the unbiased modulationSmall ER, Larger ER, Binary tree, d-regular, Karate, Dolphins, Football, and Eurosis · Frobenius norm error0.0488(9) -> 0.0437(9); 0.0503(4) -> 0.0448(4); 0.0453(4) -> 0.0410(4); 0.0490(2) -> 0.0434(2); 0.0492(6) -> 0.0439(6);HelpedGeneral Graph Random Features (ICLR 2024 · 0 citations)
Use 10 uniform chunks instead of full autoregressive codingMultiset 1k · compression rate (bits)1833 -> 2707HurtPractical Shuffle Coding (NeurIPS 2024 · 0 citations)

What reviewers will expect

Datasets and baselines named in the 20 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.

Datasets

  • TWITTER · 2 of 20 papers
  • RECIPE · 2 of 20 papers
  • CLASSIC · 2 of 20 papers
  • AMAZON · 2 of 20 papers
  • Orbit · 2 of 20 papers
  • MPEG7 · 2 of 20 papers
  • SIFT1M · 2 of 20 papers
  • GIST1M · 2 of 20 papers

Recent preprints

Loading recent arXiv preprints…

    Who publishes on it

    Institutions on the most papers of this approach since 2023 (of 72 papers; from OpenAlex and the papers' affiliations).

    Use it

    Download BibTeX of 40 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.

    Same keyword cluster

    Elsewhere in the topic

    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).

    Get the next scime Dispatch and a monthly field update

    Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy