First author of MUSE, a benchmark for large vision-language models on multimodal understanding in situated education
1
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-21) 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—
The output axis measures publication volume and productivity over time. Evolve's snapshot holds only one reviewed paper, so no sustained output count can be computed yet; this axis is left unscored rather than estimated.
Impact—
The impact axis would track citations and uptake of this researcher's work. No citation counts are present in the provided data, so the axis is reported as missing rather than guessed.
Focus100/100
The focus score is 100 because every reviewed output sits in one field: artificial intelligence (multimodal machine learning), a field distribution of a single paper at full weight. MUSE concentrates on multimodal understanding in situated education, spanning artistic image understanding, affective computing, and visual grounding. With only one reviewed paper, perfect concentration reads as early specialization rather than an established pattern.
Collaboration32/100
The collaboration score is 32, based on 4 distinct coauthors in the snapshot. The recurring circle is small and named: Xun Wei Yee, Wei Li, Mun Thye Mak, and Wee Siong Ng. On MUSE, Luyao Zhu appears as first author (non-corresponding), consistent with hands-on work inside a compact team rather than leading a large distributed consortium.
Activity—
The activity axis measures publishing cadence and recent output. With a single 2026 paper in the catalog, there is no longitudinal signal to score; the axis is left unscored.
Reproducibility55/100
Reproducibility averages 55 across one assessed paper (reproPapers = 1), and MUSE carries a medium repro-readiness rating. Evolve's review credits the annotation-first design, which should make the 2,400-question resource rebuildable, but flags the absence of uncertainty estimates and chance-level baselines across the 30 evaluated models, limiting independent verification.
Research summary
Luyao Zhu is an early-stage researcher whose reviewed work to date centers on evaluating large vision-language models. The Evolve catalog holds one first-authored paper: MUSE (arXiv, 2026), a benchmark for multi-modal understanding in situated education. MUSE organizes 12 tasks and 2,400 questions around artistic images, drawing on affective computing and visual grounding. Evolve's review names the annotation-first design as the paper's strength, while flagging missing uncertainty estimates and chance baselines and rating reproducibility readiness as medium. Radar scores reflect this narrow base: focus is 100, with all reviewed output in artificial intelligence (multimodal machine learning); collaboration is 32 across 4 distinct coauthors; reproducibility averages 55 over a single assessed paper. Output, impact, and activity are left unscored because the underlying bibliometric data is not yet sufficient. The recurring coauthor circle is small and named — Xun Wei Yee, Wei Li, Mun Thye Mak, and Wee Siong Ng — consistent with compact-team benchmark engineering rather than large-consortium work. No institution or country is recorded in the snapshot, and the profile is name-resolved from public bibliographic data tied to ORCID 0000-0002-7422-7318; conclusions are scoped to this bibliographic identity. Overall, this is a focused benchmark-builder profile: one substantive first-author artifact, strong topical concentration, and limited longitudinal signal.
Collaboration pattern
The collaboration footprint in this snapshot is compact: 4 distinct coauthors and a score of 32. The recurring circle — Xun Wei Yee, Wei Li, Mun Thye Mak, and Wee Siong Ng — suggests a stable small-team pattern, with Luyao Zhu contributing as first author on the reviewed benchmark.
Research keywords
Highlights
MUSE: a 12-task benchmark for situated education
Annotation-first evaluation design
30-model capability snapshot
Timeline
- 2026First-authored MUSE appears on arXiv: a benchmark for multi-modal understanding in situated education.
- 2026MUSE ships 12 tasks and 2,400 questions centered on artistic-image understanding and affective computing.
- 2026The benchmark evaluates 30 vision-language models, with Evolve noting missing uncertainty estimates and chance baselines.
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 scores are computed from OpenAlex-based public bibliographic data only; no curated publication list beyond the Evolve-reviewed catalog is included.
- The profile is name-resolved from public bibliographic data tied to ORCID 0000-0002-7422-7318; 'Luyao Zhu' is a common name, and institution and country are not recorded.
- Output, impact, and activity dimensions are unscored because the underlying data is insufficient; do not read the radar as a completeness measure.
- Reproducibility (55) and collaboration (32) rest on a single reviewed paper and 4 distinct coauthors, so both can shift materially as the catalog grows.