Fraunhofer FIT researcher in vision-language prompt learning, noisy-label correction and LLM alignment
2
papers here
1
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
10
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—
Evolve has not yet computed a normalized output score for this profile. OpenAlex does report an h-index of 10 and an i10-index of 10, but the scoring pipeline needs more catalogued papers. This axis would measure publication volume and productivity relative to field peers.
Impact—
No impact score is available yet: this profile lacks enough citation-linked, catalogued papers for Evolve to compute normalized citation influence. This axis would measure citation-based impact, such as citations per paper and field-normalized influence, once sufficient reviewed records accumulate.
Focus100/100
Focus scores 100/100. The OpenAlex topic profile is tightly concentrated: Topic Modeling and Research Data Management Practices lead with 24 works each, followed by Scientific Computing and Data Management (17), Library Science and Information Systems (17), and Advanced Neural Network Applications (15). The field distribution records computer vision / machine learning, matching the two 2026 catalog papers on vision-language models and LLM alignment.
Collaboration46/100
Collaboration scores 46/100, a mid-range value. The last three years show 66 works but only nine distinct coauthors, indicating a compact network. Recurring partners include Lingxiao Kong, Steffen Staab, Cong Yang, Oya Beyan, Zeyi Shao, Haowen Hua, Jiaxin Zhang, and John See. On the two 2026 catalog papers Boukhers is a middle author on TecoPrompt and last author on Evolutionary Soups, neither as corresponding author.
Activity100/100
Activity scores 100/100. The recent-3-year work count is 66, and annual counts climb steeply across the record: 2 works in 2015, 8 in 2021, 25 in 2023, 31 in 2025 — the maximum of the series — and 23 already recorded in 2026. The available data shows a sustained, high-tempo publication pattern with no sign of slowing.
Reproducibility55/100
Reproducibility scores 55/100, computed from a single assessed paper (reproPapers: 1, reproMean: 55). That paper, TecoPrompt, carries a medium reproducibility-readiness rating: Evolve's review credits strong high-noise gains from temporal-gated optimal-transport label rewriting but flags an assumption-heavy theory and unreported key thresholds. With only one paper assessed, this score is provisional rather than a stable estimate.
Research summary
Zeyd Boukhers is a researcher at the Fraunhofer Institute for Applied Information Technology (FIT) in Germany; this Evolve profile resolves from OpenAlex bibliographic records under ORCID 0000-0001-9778-9164. The snapshot records an h-index of 10 and an i10-index of 10, 66 works in the last three years, and a steeply accelerating curve: 2 recorded works in 2015, 8 in 2021, 25 in 2023, 31 in 2025, and 23 already in 2026. Recent catalogued papers sit in computer vision and machine learning: TecoPrompt (2026), a middle-authored arXiv preprint on temporal-conservative prompt learning for CLIP-style vision-language models under noisy labels, which Evolve reviewed with a medium reproducibility-readiness rating; and Evolutionary Soups (2026), a last-authored preprint on evolving mixture-of-experts for multi-objective LLM alignment, still awaiting review. The broader OpenAlex topic profile clusters around Topic Modeling and Research Data Management Practices (24 works each), Scientific Computing and Data Management (17), Library Science and Information Systems (17), and Advanced Neural Network Applications (15), producing a maximum focus score. Collaboration is comparatively compact — nine distinct coauthors over the last three years, including Lingxiao Kong, Steffen Staab, Cong Yang, Oya Beyan, and John See — reflected in a mid-range collaboration score. Output and impact dimensions are not yet computed and are left unscored.
Collaboration pattern
Collaboration is compact and stable: nine distinct coauthors across 66 works in the last three years, yielding a mid-range collaboration score (46/100). Recurring partners include Lingxiao Kong, Steffen Staab, Cong Yang, Oya Beyan, Zeyi Shao, Haowen Hua, Jiaxin Zhang, and John See. On the recent catalog papers Boukhers appears as a middle author on TecoPrompt and as last author on Evolutionary Soups, in neither case as corresponding author.
Research keywords
Highlights
TecoPrompt: temporal-conservative prompt learning for CLIP under noisy labels
Why TecoPrompt's reproducibility is rated medium
Evolutionary Soups: evolving mixture-of-experts for multi-objective LLM alignment
A 2026 arXiv preprint catalogued with last-author position, applying evolutionary search over mixture-of-experts configurations to multi-objective LLM alignment. The Evolve review is pending — no verdict, keywords or reproducibility rating are available yet.
Timeline
- 2015Earliest works in the OpenAlex record: 2 publications recorded for this year.
- 2021Output begins to climb: 8 recorded works, roughly double the annual pace of the preceding years.
- 2023Step change in productivity: 25 recorded works, the sharpest single-year jump in the series.
- 2025Series peak: 31 recorded works in a single year, the highest annual count in the provided data.
- 202623 works recorded so far; two arXiv preprints — TecoPrompt and Evolutionary Soups — enter the Evolve catalog.
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 figures come from OpenAlex only; the output and impact dimensions have not yet been computed by Evolve.
- The OpenAlex topic profile (topic modeling, research data management, library science) diverges from the computer-vision papers catalogued here, so the profile may mix similarly named records; it is name-resolved from public bibliographic data under ORCID 0000-0001-9778-9164.
- The provided countsByYear series has no recorded works for 2018, and the 2026 figure covers only part of the year.
- The reproducibility score rests on a single assessed paper; the second catalog paper (Evolutionary Soups) awaits review.