Theory-grounded scientific machine learning: neural turbulence closures and network-based retail analytics
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-01) 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—
Output metrics are not yet available. This axis would measure the volume and cadence of publications and preprints over time. The reviewed catalog currently contains only two 2026 arXiv preprints, too small and too recent a base for computing a score, so no estimate is given.
Impact—
Impact metrics are unavailable. This axis would measure citation uptake and influence within the field, such as citation counts and h-index-type aggregates. Both reviewed papers are 2026 arXiv preprints, so no citation-based measurement exists yet; no score is estimated.
Focus0/100
Focus scores 100, reflecting a field distribution concentrated in a single domain: fluid dynamics / scientific machine learning (turbulence modelling), the field of Zhang's first-authored turbulence closure paper. Notably, the reviewed catalog also includes one data-mining paper on co-purchase networks, so bibliometric attribution may under-represent that second strand; the concentration is real but rests on a very small sample.
Collaboration41/100
Collaboration scores 39, computed from 6 distinct coauthors in the record. The frequent-coauthor list is anchored by George Em Karniadakis, alongside Xiao Han, Xin Zhao, Jiechun Lei, Moxuan Zheng, and Youting Wang. Zhang is first author on the turbulence paper and a middle author on the co-purchase network paper, a pattern of leading some projects while contributing within teams. The score indicates a compact, still-forming network.
Activity—
Activity data is insufficient. This axis would measure publishing cadence across years, such as annual output and continuity of research engagement. The record shows papers dated 2026 only, so no multi-year activity pattern can be computed and no score is assigned.
Reproducibility78/100
Reproducibility scores 100, but this mean is computed from just one scored paper (reproPapers = 1): the first-authored turbulence closure paper, whose reproducibility readiness Evolve rates high. The second reviewed paper, on co-purchase networks, is rated only medium because its inference is untested and no code is released. The perfect score therefore reflects one strong case rather than a sustained track record.
Research summary
Zhen Zhang is an early-stage researcher whose reviewed output to date spans two quantitative areas: theory-grounded scientific machine learning for turbulence modelling, and network-science approaches to retail analytics. In the fluid dynamics strand, Zhang is first author of 'Well-posedness of neural turbulence closures and tangent dissipation' (2026, arXiv), which proves a well-posedness theorem for coupled neural turbulence closures using monotone operator arguments and tangent dissipation, supported by channel-flow experiments in an RANS/eddy-viscosity setting. Evolve judged the work sound but flagged an unverified global margin, and rated reproducibility readiness high. In a second strand, Zhang is a middle author on 'Support Thresholds, Not Algorithms, Limit Rare-Association Recovery in Co-Purchase Networks' (2026, arXiv), which argues that support thresholds, rather than mining algorithms such as Apriori, block recovery of rare associations, and shows noise-corrected filtering recovering 80–100% of them with persistent significance; Evolve rated it plausible and well-designed but in need of revision, with medium reproducibility readiness. The collaboration footprint is compact: 6 distinct coauthors, including George Em Karniadakis, Xiao Han, Xin Zhao, Jiechun Lei, Moxuan Zheng, and Youting Wang. Radar scores show focus 100 and reproducibility 100 (the latter from a single scored paper), with collaboration at 39; output, impact, and activity lack sufficient data. Because 'Zhen Zhang' is a common name, this profile is name-resolved from public bibliographic data.
Collaboration pattern
The bibliometric record lists 6 distinct coauthors, with George Em Karniadakis, Xiao Han, Xin Zhao, Jiechun Lei, Moxuan Zheng, and Youting Wang as frequent collaborators. Zhang is first author on the turbulence closure paper and a middle author on the co-purchase network paper, consistent with leading a theory-driven project and contributing within a team elsewhere. No institutional affiliation is provided. The collaboration radar score of 39 fits a compact, still-forming coauthor pool.
Research keywords
Highlights
Well-posedness theorem for coupled neural turbulence closures
Support thresholds, not algorithms, limit rare-association recovery
Timeline
- 2026First-author arXiv preprint on well-posedness of neural turbulence closures and tangent dissipation, with channel-flow experiments.
- 2026Middle-author arXiv preprint on support thresholds and rare-association recovery in co-purchase networks.
- 2026Bibliometric record shows 6 distinct coauthors, including George Em Karniadakis.
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-derived public data only; no ORCID, institution, or country was available to verify identity.
- 'Zhen Zhang' is a common name; this profile is name-resolved from public bibliographic data and may merge or split same-name researchers.
- Output, impact, and activity dimensions lack sufficient data: both reviewed papers are 2026 arXiv preprints, too recent for citation-based measurement.
- The reproducibility score of 100 is derived from a single scored paper; the second reviewed paper reports no code and is rated medium for reproducibility.