University of Chicago statistician advancing conformal prediction, model calibration, and algorithmic stability theory.
3
papers here
3
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
26
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—
No output score has been computed yet. This axis would measure publication volume and productivity over time; the snapshot does record 47 works in the last three years and annual counts from 2013 onward, but Evolve has not yet converted that evidence into a score.
Impact—
No impact score is available yet. This axis would measure citation-based influence such as citation counts and field-normalized rates; the snapshot lists h-index 26 and i10-index 40, but these raw figures have not been turned into an Evolve impact score.
Focus0/100
The focus score is 0, and the field distribution explains why: the snapshot splits evenly into thirds — conformal prediction, model calibration and evaluation, and statistical learning theory (0.33 each). Topic counts confirm a deliberately broad portfolio rather than concentration, spanning statistical methods and inference (88 works), machine learning topics (22 and 21), and clinical-trial statistics (19).
Collaboration32/100
The collaboration score is 32, based on a snapshot recording only 4 distinct coauthors, so the radar reflects limited measured breadth rather than an isolated working style. The frequent-coauthor list names Chen Cheng, Ruiting Liang, Anirban Chatterjee, and Richard J. Samworth, and on all three Evolve-reviewed papers Barber holds a senior position — first author on one, last author on the other two.
Activity100/100
Activity scores a perfect 100. The OpenAlex counts-by-year record runs unbroken from 2013 (1 work) through 2026, rising to 12 works in 2019, 13 in 2020, and a snapshot-high 28 in 2026, with 47 works in the last three years — one of the most sustained output trajectories the radar measures.
Reproducibility85/100
Reproducibility scores 85, a mean over the 3 papers Evolve has reviewed. The rolling conformal prediction and rankECE calibration papers are rated high readiness — theory-forward with reusable procedures — while the ensembling-stability paper is rated medium. Candidly stated limitations (the factor-2 coverage gap, bounded-output assumptions) support a high score rather than detract from it.
Research summary
Rina Foygel Barber is a statistician at the University of Chicago whose recorded work, resolved via ORCID 0000-0002-4854-8644 in an OpenAlex snapshot, sits overwhelmingly in statistical methods and inference (88 topic-assigned works), with further clusters in probability and statistical research (37), machine learning and data classification (22), machine learning and algorithms (21), and statistical methods in clinical trials (19). The three preprints Evolve has reviewed — all 2026 arXiv postings — share one style: rigorous mathematical theory brought to bear on practical statistical machine learning questions. 'Rolling Conformal Prediction in Sequential Model Training' (last author) gives a calibrate-then-train procedure for streams holding 1−2α marginal coverage without stability assumptions; 'A Ranking Approach for Measuring Calibration' (last author) introduces rankECE, a tuning-free, rank-based miscalibration measure; 'Algorithmic Stability via Ensembling' (first author) proves that averaging over an ensemble stabilizes arbitrary algorithms against data perturbations. Two of the three carry high reproducibility readiness. Activity metrics are the strongest signal: 47 works in the last three years, an activity score of 100, and unbroken annual output from 2013 through 2026, with h-index 26 and i10-index 40 recorded in the snapshot. Output and impact dimension scores remain uncomputed.
Collaboration pattern
Across the Evolve-reviewed work Barber occupies senior authorship — first author on 'Algorithmic Stability via Ensembling' and last author on the two other 2026 preprints — a pattern consistent with leading a group. The snapshot lists only 4 distinct coauthors, naming Chen Cheng, Ruiting Liang, Anirban Chatterjee, and Richard J. Samworth as frequent collaborators; that small count likely reflects incomplete coauthor coverage in the OpenAlex snapshot rather than a solitary style.
Research keywords
Highlights
Rolling conformal prediction for sequential model training
rankECE: a tuning-free, rank-based calibration measure
Ensembling as a general algorithmic stabilizer
Timeline
- 2013Earliest year in the OpenAlex counts-by-year record, with 1 work recorded.
- 2019Annual output reaches 12 works, then 13 in 2020 — the steepest early rise in the record.
- 202314 works recorded, the strongest full year in the snapshot to that point.
- 202628 works recorded in the snapshot so far — the highest single-year count — including all three Evolve-reviewed arXiv preprints: rolling conformal prediction, rankECE calibration, and algorithmic stability via ensembling.
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.
- All bibliometrics and dimension scores derive from an OpenAlex snapshot (identity matched via ORCID 0000-0002-4854-8644); the 4 distinct coauthors likely undercount the real collaboration network.
- Output and impact dimension scores are not yet computed; h-index 26 and i10-index 40 are quoted directly from the snapshot and are not Evolve scores.
- Qualitative paper assessments cover only the three Evolve-reviewed 2026 arXiv preprints; earlier work is represented by bibliometric counts alone.
- The 2026 annual count (28) reflects a snapshot in progress and may change as the year's publications are indexed.