NLP researcher at Masaryk University: topic modeling, prompt-robust language models, and AI in education
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. This axis would measure publication volume and scale of research output relative to field norms; the snapshot lacks the normalized output metrics needed, so only raw yearly counts exist and no estimate is offered.
Impact—
Not yet scored. This axis would measure citation uptake and downstream influence of the work, beyond the raw h-index of 5 and i10-index of 1; the snapshot lacks sufficient citation data, so no score or estimate is made.
Focus100/100
Focus is maximal (100/100). Of 125 topic-assigned works, topic modeling (48) and NLP techniques (44) account for 92 (~74%), with smaller threads in domain adaptation and few-shot learning (16), adversarial robustness (10), and information retrieval and search behavior (7); the field distribution lists a single field, NLP (AI in education). This indicates a tightly concentrated research program.
Collaboration41/100
Collaboration scores 41/100. The OpenAlex record lists 7 distinct coauthors in the recent window, and the recurring circle is a compact Masaryk University group — Jan Nehyba, Jirina Karasova, Martin Fico, Lucie Škarková, Markéta Košatková, David Kosatka, and Frederic Sadrieh. The pattern suggests small-team, locally anchored projects rather than large cross-institution consortia; both catalog papers fit this, spanning first-author and last-author roles.
Activity97/100
Activity is near-ceiling (97/100). The record shows 27 works in the last three years and a clear acceleration: from 1 work in 2016 and 2 in 2019 to 8 (2021), 8 (2022), 9 (2023), 6 (2024), 13 (2025), and 8 already in 2026. Output has run at roughly 6-13 works per year since 2021, indicating a sustained, high-tempo publishing cadence.
Reproducibility55/100
Reproducibility scores 55/100, based on a single assessed paper (reproMean 55 across 1 paper), so the estimate is thin. Evolve rated both catalog papers at medium reproducibility readiness: the Edustories dataset paper and the prompt-robustness comparison. Reviewer notes flag missing statistical rigor in the LLM-versus-expert analysis and an under-tuned baseline (COIN), suggesting replicable artifacts but analysis gaps.
Research summary
Michal Štefánik is a natural language processing researcher at Masaryk University whose OpenAlex-resolved record shows a compact but rapidly accelerating publication trajectory. Annual output rises from single works before 2021 (1 in 2016, 2 in 2019, 4 in 2020) to a sustained cadence of 6-13 per year from 2021 onward — 8 in 2021, 9 in 2023, 13 in 2025, and 8 already in 2026 — totaling 27 works in the last three years, which drives an activity score of 97. Topical concentration is maximal (focus 100): topic modeling (48) and NLP techniques (44) dominate, supported by domain adaptation and few-shot learning (16), adversarial robustness (10), and information retrieval and search behavior (7). Two 2026 papers reviewed by Evolve extend this program into AI in education and robustness evaluation: the first-authored Edustories dataset of 1,492 teacher-written classroom cases, and a last-authored controlled comparison of prompt-robust training strategies. Citation impact remains early-stage (h-index 5, i10-index 1), so output and impact are not scored; collaboration is compact (41) with 7 distinct coauthors, and reproducibility is 55 on one assessed paper.
Collaboration pattern
The record shows a compact collaboration footprint: 7 distinct coauthors in the recent window (collaboration score 41/100), centered on a recurring Masaryk University group — Jan Nehyba, Jirina Karasova, Martin Fico, Lucie Škarková, Markéta Košatková, David Kosatka, and Frederic Sadrieh. The 2026 catalog papers fit this pattern: a first-authored dataset effort and a last-authored study, suggesting a small core team where Štefánik both leads data work and supervises junior collaborators.
Research keywords
Highlights
Edustories: 1,492 teacher-written classroom cases for AI-in-education research
Simple batch schedules rival robustness losses for prompt robustness
Dual roles in 2026: leading dataset building and supervising evaluation studies
Across the two 2026 releases, Štefánik is first author on the education dataset and last author on the robustness study — a pairing that fits the record's compact 7-coauthor circle and suggests a shift toward leading data-collection efforts while directing junior collaborators' analyses.
Timeline
- 2016First work appears in the OpenAlex record (1 publication).
- 2021Output steps up to 8 works, beginning a sustained multi-year publishing cadence.
- 20239 works recorded, consolidating the topic-modeling and NLP core of the research program.
- 2025Most productive year in the record so far: 13 works.
- 2026Two Evolve-reviewed releases — the Edustories dataset (first author) and the prompt-robustness study (last author) — appear among 8 works so far.
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, topic counts, and collaboration figures are computed from OpenAlex only; citation coverage may lag published work.
- Output and impact dimensions are unscored in this snapshot for insufficient data; h-index (5) and i10-index (1) are quoted as raw values.
- The reproducibility score rests on a single assessed paper, so read it as indicative rather than settled.
- Identity is name-resolved from public bibliographic data (OpenAlex, ORCID 0000-0003-1766-5538); the listed country code (JP) does not obviously match the stated affiliation (Masaryk University) and should be verified.