Multimodal AI researcher bridging brain tumor diagnosis models and vision-language benchmarks (OpenAlex profile)
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-19) 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—
This axis measures total publication volume and cumulative productivity. The underlying data is not yet available: the OpenAlex record shows an h-index of 0 and only 2 indexed works, too thin to score output reliably.
Impact—
This axis measures citation-based influence, such as citations per publication and field-weighted impact. Citation aggregates are not yet available for this profile — h-index and i10-index both stand at 0 — so no score is assigned.
Focus0/100
The focus score of 0 is derived from an unusually even field distribution: OpenAlex assigns 0.5 to medical imaging AI (neuroradiology) and 0.5 to computer vision (multimodal machine learning), and each of five indexed topics — from glioma diagnosis to few-shot learning — carries just one work. This reads as a deliberate two-sided portfolio pairing clinical MRI applications with vision-language benchmarking, rather than scattered or inactive research.
Collaboration74/100
A score of 74 rests on a wide network: 39 distinct coauthors are recorded, and both catalog papers are multi-author team efforts in which Wang holds a middle, non-corresponding position. Recurring collaborators include Yinong Wang (Joyce), Jianwen Chen (Joyce), Zhou Chen (Joyce), Shuwen Kuang (Joyce), Haoning Jiang (Joyce), plus Bing Wang, Lei Wu, Bin Tang, and Li Meng, mixing clinical and computer-vision partners.
Activity32/100
Activity scores 32 on a thin but recent record: the counts-by-year series contains only 2026 with 2 works, and the recent three-year window likewise totals 2 works. Both entries are arXiv preprints appearing in the same year, so the low score reflects a short observable history rather than a slowdown — more time and additional indexed papers are needed before the trend is meaningful.
Reproducibility55/100
Evolve's reproducibility review averages 55 across exactly 2 papers, and both catalog entries carry a medium reproducibility readiness rating. The brain tumor foundation model paper shows convincing broad validation but internal inconsistencies, while the video benchmark is well designed yet statistically limited by small datasets, including a 20-video CMD set. The mid-range score reflects convincing validation alongside noted inconsistencies and dataset-size limits.
Research summary
Wang is a researcher whose indexed record is at an early stage, sitting at the interface of clinical neuro-oncology AI and multimodal machine learning. OpenAlex resolves the profile through two 2026 arXiv preprints: a vision-language foundation model for comprehensive brain tumor diagnosis from preoperative multimodal data, and a benchmark study arguing that visual question answering for large video models has drifted toward text shortcuts. Together they anchor an even 50/50 field split between medical imaging AI (neuroradiology) and computer vision (multimodal machine learning), which also explains the focus score of 0 — the portfolio is deliberately two-sided rather than narrow. Collaboration is the strongest dimension (74): 39 distinct coauthors appear in the network, with recurring partners such as Yinong Wang (Joyce), Jianwen Chen (Joyce), and Haoning Jiang (Joyce), and Wang contributes as a middle author on both catalog papers. Activity is still thin (32), with 2 works in the recent three-year window, both dated 2026; citation aggregates (h-index 0, i10-index 0) are not yet meaningful, so output and impact are left unscored. Evolve's reproducibility review averages 55 across two papers, each rated medium readiness.
Collaboration pattern
Wang works in team-based settings: both catalog papers are multi-author efforts with a middle, non-corresponding author position, and the OpenAlex network records 39 distinct coauthors (collaboration score 74). Recurring names include Yinong Wang (Joyce), Jianwen Chen (Joyce), Zhou Chen (Joyce), Shuwen Kuang (Joyce), and Haoning Jiang (Joyce), alongside Bing Wang, Lei Wu, Bin Tang, and Li Meng — a mix consistent with the 50/50 split between clinical imaging AI and computer vision.
Research keywords
Highlights
Vision-language foundation model for radiologist-comparable brain tumor MRI diagnosis
This 2026 arXiv preprint presents a vision-language foundation model that diagnoses 12 brain tumor types from preoperative multimodal data under the WHO CNS5 framework, with calibrated uncertainty. Evolve finds the broad validation convincing, though internal inconsistencies are noted; reproducibility readiness is rated medium.
Benchmark puts the “visual” back in video question answering
In this 2026 benchmark study, frontier video models collapse when multiple-choice options become visual frames — Gemini drops from 74% to 30%. Evolve credits the compelling design but flags small datasets, especially the 20-video CMD set, as limiting the statistics; reproducibility readiness is rated medium.
Timeline
- 2026Co-authored an arXiv preprint on a vision-language foundation model for precise and comprehensive brain tumor diagnosis from preoperative multimodal data, in a middle author position.
- 2026Co-authored the arXiv benchmark study “Not Another Text Benchmark,” which restores visual grounding to visual question answering for large video models.
- 2026Bibliometric snapshot: 2 works indexed in 2026, 39 distinct coauthors, and five topics spanning glioma diagnosis, few-shot learning, saliency detection, multimodal ML, and radiomics.
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.
- Identity is name-resolved from public OpenAlex bibliographic data; only the surname “Wang” is available, with no institution, country, or ORCID, so author disambiguation carries uncertainty.
- All bibliometrics come from OpenAlex only; h-index and i10-index are recorded as 0 and citation aggregates are absent, so output and impact are left unscored rather than estimated.
- Evidence rests on 2 arXiv preprints (both 2026); no peer-reviewed venues, funding, or career history are in scope.
- The reproducibility mean of 55 is averaged over only 2 papers, so the score is indicative rather than stable.