Graph ML researcher at Vanderbilt: formal verification of GNNs and LLM-driven autonomous knowledge discovery
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
27
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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—
Not yet scored: Evolve's output axis measures publication volume and venue-normalized productivity over time, but this snapshot lacks the complete per-year works record needed to compute it, so no estimate is offered here.
Impact—
Not yet scored: the impact axis measures citation-based influence such as field-normalized citation rates. The snapshot reports raw h-index 27 and i10-index 47, but lacks the full citation record Evolve needs, so no score is given.
Focus0/100
The focus score sits at the floor of the 0-100 scale, and the raw field distribution explains why: the recent split is exactly even — 0.5 for machine-learning formal verification of GNNs in power systems and 0.5 for LLM-based autonomous scientific discovery. Underlying topics are similarly spread across Advanced Graph Neural Networks (51), Complex Network Analysis (45), recommender systems (30), topic modeling (24), and graph theory and algorithms (24).
Collaboration53/100
The collaboration score of 53 is mid-range, grounded in 13 distinct coauthors in the snapshot. The recurring roster splits into two clusters matching the field split: Taylor T. Johnson, Diego Manzanas Lopez, Ben Wooding, Anne M. Tumlin, and Zhenxuan Shao around verification and graph work, plus Qinwen Ge, Bo Ni, Haowei Fu, and Ngoc N. Tran. Both reviewed 2026 papers are multi-author efforts, with Derr in middle and last author positions and neither marked corresponding.
Activity100/100
The activity score is the maximum 100, and the yearly counts bear it out: indexed works rise from 3 (2017), 5 (2018), and 7 (2019) to 13 (2020), then hold between 10 and 12 through 2023 before accelerating to 19 (2024), 21 (2025), and 35 (2026). The snapshot records 75 works in the last three years, so current output is both frequent and still rising.
Reproducibility78/100
Reproducibility scores 78, computed as the mean over the 2 papers Evolve has reviewed. The reachability-based GNN verification paper is rated high readiness — Evolve finds the GraphStar-set method sound with a strong CORA comparison — while AutoKD is rated medium, its evaluation described as small and rater-dependent. With only 2 reviewed papers, the score is a thin, early-stage signal.
Research summary
Tyler Derr is a researcher at Vanderbilt University (US) whose profile, name-resolved from public OpenAlex bibliographic data, centers on graph machine learning and its extension into two directions captured by Evolve's 2026 catalog reviews: reachability-based formal verification of graph neural networks aimed at power systems, and LLM-driven autonomous knowledge discovery (AutoKD). The bibliometric snapshot reports an h-index of 27, an i10-index of 47, and 75 works over the last three years, with annual output climbing from 3 indexed works in 2017 to 35 in 2026 — the peak of a near-monotonic rise that anchors the maximal activity score. Topic mass is concentrated in Advanced Graph Neural Networks (51) and Complex Network Analysis Techniques (45), with supporting clusters in recommender systems (30), topic modeling (24), and graph theory and algorithms (24). The recent field distribution splits exactly even (0.5/0.5) between GNN formal verification and LLM-based scientific discovery, which explains the minimal focus score and signals a deliberate pivot. The two reviewed papers yield a mean reproducibility of 78, with the verification work rated high. Collaboration is mid-range (13 distinct coauthors), anchored by repeat partners including Taylor T. Johnson and Diego Manzanas Lopez.
Collaboration pattern
Derr's collaboration footprint in the snapshot spans 13 distinct coauthors, a mid-range breadth (score 53). Recurring names form two clusters aligned with his current field split: a formal-methods cluster with Taylor T. Johnson and Diego Manzanas Lopez, alongside Ben Wooding, Anne M. Tumlin, and Zhenxuan Shao; and a graph/data-mining cluster with Qinwen Ge, Bo Ni, Haowei Fu, and Ngoc N. Tran. Both 2026 catalog papers are coauthored efforts, with Derr in middle and last author positions.
Research keywords
Highlights
First edge-aware formal verification of GNNs via GraphStar sets
AutoKD: LLM discovery loop with a persistent insight graph
Timeline
- 2017Earliest works in the snapshot: 3 indexed publications as the bibliographic record begins.
- 2020Output nearly doubles year over year, reaching 13 indexed works from 7 in 2019.
- 2024The climb steepens: 19 indexed works, up from 11-12 per year across 2022-2023.
- 202635 indexed works — the highest annual count in the 2017-2026 series; the recent three-year total stands at 75.
- 2026Two arXiv preprints enter the Evolve catalog: edge-aware formal verification of GNNs and AutoKD autonomous knowledge discovery.
Papers in the Evolve catalog
Frequent co-authors
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 indicators are name-resolved from an OpenAlex-based snapshot and may not capture every publication, venue, or name variant.
- Output and impact radar axes are unscored; only raw h-index (27) and i10-index (47) are available, without the full citation record.
- The reproducibility mean (78) rests on only 2 Evolve-reviewed papers, both 2026 arXiv preprints — a thin, early-stage sample.
- No verified biography, award, or career-history data was provided; the profile is limited to bibliometrics and the reviewed papers.