Evolve.Pub
Sign inScan your paper
CF

Chaoqun Fei

South China Normal University · China

Unclaimed · AI-structured profileORCID OpenAlex

Graph ML researcher, South China Normal University — curvature-guided GNN sampling and Ricci-flow graph coarsening.

2

papers here

2

reviewed by Evolve

2

first-author

0

corresponding

—

works · OpenAlex

6

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-15) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output: insufficient dataOutputImpact: insufficient dataImpactFocus: 0/100FocusCollaboration: 39/100CollaborationActivity: 79/100ActivityReproducibility: 55/100Reproducibility

Output—

Not yet scored: the underlying publication-volume data is insufficient. This axis would measure cumulative output and document types across the career; only partial per-year counts and a 14-work three-year total are available here.

Impact—

Not yet scored: citation-based impact data is insufficient. Beyond headline counts such as an h-index of 6 and an i10-index of 3, the snapshot lacks the citation distribution and field-normalized comparisons this axis measures.

Focus0/100

The radar focus score is 0, but the raw field attribution is narrow: 50% graph machine learning and 50% graph machine learning (geometric deep learning). Topic counts add texture: Advanced Graph Neural Networks leads with 18, followed by Semantic Web and Ontologies (9), Topological and Geometric Data Analysis (8), Graph Theory and Algorithms (7), and Topic Modeling (6).

Collaboration39/100

Collaboration scores 39, reflecting a compact network of 6 distinct coauthors in the snapshot. Recurring collaborators include Tinglve Zhou, Guoxuan Li, Chuanqing Wang, Yangyang Li, and Tianyong Hao. Both reviewed 2026 preprints are first-authored without corresponding-author flagging, consistent with a small-team working mode; the low distinct-coauthor count also limits the cross-institution reach visible in this data.

Activity79/100

Activity scores 79, driven by a clear acceleration: counts-by-year show 1–2 works per year from 2017 to 2022, then 6 in 2023, 5 in 2025, and 9 in 2026, with 14 works in the last three years. Note the series has no 2024 entry and 2026 is likely still filling in. The recent concentration is consistent with the two 2026 arXiv preprints under review.

Reproducibility55/100

Reproducibility scores 55, the mean across the 2 papers Evolve has reviewed, both rated medium readiness. ERC-LG shows credible scaling results, but accuracy wins are mostly within noise; RicciPool has thorough ablations, yet wins lack significance tests and reported numbers conflict. Consistent significance testing would be the clearest path to a higher score.

Research summary

Chaoqun Fei is a graph machine learning researcher at South China Normal University (ORCID 0000-0001-9164-0019), name-resolved from OpenAlex bibliographic records. The citation record is early-stage: an h-index of 6 and an i10-index of 3, against 14 works in the last three years. Output is accelerating — counts-by-year rise from 1–2 works annually (2017–2022) to 6 in 2023, 5 in 2025, and 9 in 2026. Field attribution splits evenly between graph machine learning and graph machine learning (geometric deep learning); topic counts are led by Advanced Graph Neural Networks (18), followed by Semantic Web and Ontologies (9), Topological and Geometric Data Analysis (8), Graph Theory and Algorithms (7), and Topic Modeling (6). Evolve has reviewed two 2026 first-authored arXiv preprints: ERC-LG, which uses Johnson-Lindenstrauss projections and batched conjugate-gradient solves for curvature-guided subgraph sampling, and RicciPool, which adds Ollivier-Ricci flow edge-reweighting to graph pooling. Both carry medium reproducibility readiness, averaging 55 across the two reviewed papers; the verdicts credit credible scaling results and thorough ablations while flagging accuracy wins within noise and missing significance tests. Collaboration is compact — 6 distinct coauthors, including Tinglve Zhou, Guoxuan Li, Chuanqing Wang, Yangyang Li, and Tianyong Hao.

Collaboration pattern

Collaboration is compact: 6 distinct coauthors, with Tinglve Zhou, Guoxuan Li, Chuanqing Wang, Yangyang Li, and Tianyong Hao recurring most often. Both reviewed 2026 preprints are first-authored and not flagged as corresponding author, suggesting a small, stable team pattern rather than large cross-institution consortia. Radar collaboration score: 39.

Research keywords

graph neural networkssubgraph samplingeffective resistance curvaturediscrete graph curvatureOllivier-Ricci curvatureRicci flowgraph poolinggraph coarseningJohnson-Lindenstrauss projectionspectral clusteringtopological and geometric data analysissemantic web and ontologies

Highlights

Timeline

  1. 2017First year in the provided counts-by-year series, with 1 recorded work — the start of a flat 2017–2022 stretch of 1–2 works per year.
  2. 2023Output steps up to 6 works in the year, the first clear acceleration visible in the series.
  3. 20255 works recorded; the recent three-year window now totals 14 works.
  4. 2026Series peak with 9 works; two first-authored arXiv preprints (ERC-LG, RicciPool) enter the Evolve catalog, both rated medium reproducibility readiness.

Papers in the Evolve catalog

Frequent co-authors

Tinglve Zhou×2Guoxuan Li×1Chuanqing Wang×1Yangyang Li×1Tianyong Hao×1Yangyang Li×1

Is this you? Claim this profile

Enter the email you published with. We'll send a one-time claim link; after claiming you can edit the introduction or hide this page. Requests from a mismatched address go to manual review.

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

  • Bibliometric snapshot sourced from OpenAlex (identity: openalex); the output and impact radar axes are unscored because their underlying data is missing.
  • Profile is name-resolved from public bibliographic data via ORCID 0000-0001-9164-0019; similar names can collide across OpenAlex records — verify before reuse.
  • Only the two 2026 arXiv preprints carry Evolve review verdicts and reproducibility scores; the remaining works are counted but not independently reviewed.
  • The counts-by-year series has no 2024 entry and includes a partial 2026, so trend readings should be treated as provisional.
Chaoqun Fei — Researcher Profile | Evolve.Pub