University of Michigan theorist advancing logconcave sampling, Markov chain mixing, and optimization complexity.
2
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
5
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) 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 — the underlying data (complete indexed works and publication-venue records) is insufficient. This axis would measure publication volume and productivity over time, here only partially visible as 19 recent three-year works.
Impact—
Not yet scored — citation-level data is insufficient for this axis. It would measure citation uptake and scholarly influence of the work; the cataloged 2026 papers are too recent for stable citation signals, so no impact estimate is offered.
Focus100/100
Focus scores 100, the axis maximum. OpenAlex field attribution places the profile in a single field — theoretical computer science / convex geometry and randomized algorithms — and topic counts concentrate on one agenda: Markov chains and Monte Carlo methods (18), point processes and geometric inequalities (14), and stochastic gradient optimization (11), with graph-algorithm topics (9 and 4) clearly secondary. Both cataloged papers sit squarely in sampling complexity and mixing analysis.
Collaboration22/100
Collaboration scores 22, reflecting a compact co-author network: only 2 distinct coauthors appear in the bibliometric data, with Yuansi Chen and Santosh S. Vempala as repeat collaborators. The cataloged pattern spans both extremes of the author list — first author on the Gaussian-cooling paper and last author on the Dikin-walks note — suggesting an individual-theory working style that develops ideas with one or two close partners rather than large teams.
Activity87/100
Activity scores 87 on a clear upward trajectory: yearly output climbed from 2 works in 2020 to 3 in 2022, then 6 in 2024, 6 in 2025, and 7 in 2026, with 19 works in the last three years alone. Both cataloged 2026 papers are arXiv preprints, including a first-authored contribution, indicating a fast-moving, highly productive current research pace.
Reproducibility20/100
Reproducibility scores 20, computed from 1 assessed paper with mean readiness 20 — Evolve rates the first-authored Gaussian-cooling paper 'low'. As a pure-theory line of work, its contributions are proof-based rather than artifact-based, and the paper leans on sampler guarantees from an unpublished companion preprint, so independent verification currently rests on reproducing the mathematics rather than released code.
Research summary
Yunbum Kook is a theoretical computer scientist at the University of Michigan (US) whose work centers on the complexity of logconcave sampling and the mixing analysis of Markov chains. The Evolve catalog includes two 2026 arXiv papers: a first-authored study of thin-shell stability along the Gaussian cooling path, which upgrades cold-start logconcave sampling from n^2.75 to nearly n^2.5 query complexity — matching the abstract Speedy walk — though the gain is isotropic-only and rests on sampler guarantees quoted from an unpublished companion preprint; and a last-authored note giving two proofs of d^2 mixing for weighted Dikin walks. Bibliometrically the profile is early-stage and name-resolved from OpenAlex: h-index 5, i10-index 4, and 19 works in the last three years, with yearly output rising from 2 works in 2020 to 7 in 2026. Topic mass concentrates in Markov chains and Monte Carlo methods (18), point processes and geometric inequalities (14), and stochastic gradient optimization (11), with graph-algorithm topics (9 and 4) secondary. Collaboration is narrow — 2 distinct coauthors, with repeat work alongside Yuansi Chen and Santosh S. Vempala. The output and impact radar axes are not yet scored.
Collaboration pattern
The co-author network is deliberately small: 2 distinct coauthors in the OpenAlex-derived data, with Yuansi Chen and Santosh S. Vempala as recurring partners. Author placement spans both ends of the list — first author on the Gaussian-cooling sampling paper and last author on the Dikin-walks proofs — a pattern typical of theory groups where senior collaborators anchor projects the profile leads, and vice versa. No large-team or experimental collaborations appear in the current data.
Research keywords
Highlights
Thin-shell stability upgrades cold-start logconcave sampling to nearly n^2.5 queries
Two proofs of d^2 mixing for weighted Dikin walks
Last-authored 2026 arXiv note establishing d^2 mixing for weighted Dikin walks via two independent proofs, extending the geometric-random-walk toolkit for sampling analysis. Evolve has not yet reviewed this paper, so no verdict, keywords, or reproducibility assessment is available.
Timeline
- 2020Earliest year in the bibliometric window: 2 indexed works recorded, the baseline of a steady climb in annual output.
- 20223 indexed works recorded, the mid-window mark before the 2024-2026 acceleration.
- 2024Output reaches 6 indexed works, opening the high-activity phase that accumulates 19 works across the last three years.
- 2026Most productive year at 7 indexed works; both Evolve-cataloged arXiv papers appear — first-authored thin-shell stability of Gaussian cooling and last-authored two proofs of d^2 mixing of weighted Dikin walks.
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 counts, h-index (5), i10-index (4), topics, and yearly output derive from OpenAlex; identity is name-resolved from public bibliographic data.
- The output and impact radar axes are unscored: indexed citation and venue data are insufficient, and no citation numbers should be inferred.
- Evolve has reviewed only 2 papers; the second has no verdict, keywords, or reproducibility assessment, and its field is unclassified.
- The reproducibility score rests on a single assessed paper (reproMean 20 from 1 paper) and should not be read as a stable estimate.