Preslav Nakov
Mohamed bin Zayed University of Artificial Intelligence · Denmark
NLP researcher at MBZUAI focused on LLM evaluation, open question answering, and Arabic-language benchmarks
3
papers here
0
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
2
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—
Insufficient data to score: this axis measures publication volume and productivity over time. The OpenAlex output fields (h-index 2, i10-index 0) conflict with the 73 recent works, suggesting a fragmentary record, so no score is computed.
Impact—
Insufficient data to score: this axis measures citation-based influence such as citation counts and h-index trajectory. The recorded figures (h-index 2, i10-index 0) appear unreliable against the work count, so the axis is left unscored.
Focus—
Insufficient data to score: this axis measures how concentrated versus diversified the research topics are. Topic labels exist (topic modeling 53, NLP techniques 34), but the underlying record is too incomplete to compute a reliable focus score.
Collaboration72/100
The record lists 36 distinct coauthors in the recent three-year window, anchored by a recurring core of twelve frequent collaborators including Muhra Almahri, Ahmed Heakl, Saeed Almheiri, Elitsa Yotkova, Violeta Kastreva, Ivan Koychev, and Dimitar Dimitrov. Across the three Evolve-reviewed papers he appears as last author once and middle author twice, never as corresponding author — a pattern of senior, distributed team science on shared benchmarks. Collaboration scores 72/100 on this evidence.
Activity100/100
Activity scores 100/100, the maximum. The OpenAlex record shows 73 works in the recent three-year window, ramping from 1 indexed work in 2024 to 4 in 2025 and 68 in 2026, the year all three Evolve-reviewed papers appeared. The curve signals an intensive, arXiv-first publishing phase, though fragmentary identity indexing means early-year counts may understate true output.
Reproducibility—
Insufficient data to score: this axis measures code, data, and artifact release practices across outputs. Evolve's reviews rate two catalog papers 'high' and one 'medium' for repro readiness, but that three-paper base is too small to score the axis.
Research summary
Preslav Nakov is profiled here from OpenAlex bibliometric data and three Evolve-reviewed 2026 arXiv papers, all from the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI). The record shows a sharply accelerating publication curve: 1 indexed work in 2024, 4 in 2025, and 68 in 2026 — 73 works across the recent three-year window — with 36 distinct coauthors. Topic frequencies place the record at the intersection of topic modeling (53 assignments), general NLP techniques (34), and three evenly weighted clusters of 14 each: misinformation, adversarial robustness, and multimodal machine learning. The Evolve-reviewed papers clarify the current focus: a semantic-correctness framework and CAP metric for open question-answering evaluation (senior/last author); EDRAC, the first naturally spoken five-dialect Arabic generative-QA benchmark with 499 passages and 4,977 QA pairs; and an LLM-evaluation study showing cultural cues steering models toward Islamic-finance framing they answer wrongly. Across these, he sits in last or middle authorship, none as corresponding author, consistent with a senior collaborative role. A major caveat: the recorded h-index (2) and i10-index (0) are incompatible with 73 recent works, so the OpenAlex identity is likely fragmented and citation-based scores should not be read as career totals.
Collaboration pattern
Thirty-six distinct coauthors appear in the recent three-year window, with a recurring core of twelve frequent collaborators spanning the reviewed papers: Muhra Almahri, Elitsa Yotkova, Violeta Kastreva, Petar Velkov, Hristo Boyanov, Dimitar Dimitrov, Ivan Koychev, Rania Elbadry, Ahmed Heakl, Saeed Almheiri, Fan Zhang, and Xueqing Peng. On the three Evolve-reviewed papers Nakov is last author once and middle author twice, never corresponding — a senior-collaborator pattern of supervising multiple small benchmark teams in parallel.
Research keywords
Highlights
Semantic correctness framework for open QA evaluation
EDRAC: first spoken five-dialect Arabic reading-comprehension benchmark
Cultural cues steer LLM answers in Islamic-finance MCQs
Timeline
- 2024First work recorded in the OpenAlex activity window: 1 indexed publication.
- 20254 works recorded in OpenAlex, a modest early ramp-up.
- 202668 works recorded, including all three Evolve-reviewed papers: the CAP open-QA evaluation framework, the EDRAC Arabic-dialect QA benchmark, and the cultural-cues LLM evaluation study.
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.
- Bibliometrics come from OpenAlex only; the recorded h-index (2) and i10-index (0) conflict with 73 recent works, so the identity record is likely fragmented and citation scores should not be read as career totals.
- Output, impact, focus, and reproducibility axes are unscored because the underlying data is insufficient.
- Qualitative judgments rest on only three Evolve-reviewed papers, all 2026 arXiv preprints not tied to peer-reviewed venues in this data.
- Identity is name-resolved from public bibliographic data (no ORCID supplied); the source lists country as 'DK', which may not match the institution — treat identity fields as provisional.