SciML researcher at RPI advancing neural operators and reliable neural PDE surrogates for computational fluid dynamics
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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 is not scored: the OpenAlex snapshot reports an h-index of 0, an i10-index of 0, and only 3 works in the last three years, too sparse to measure productivity reliably. This axis would normally capture publication volume and its growth over time.
Impact—
Impact is not scored: with h-index 0 and i10-index 0 in the current snapshot, citation-based influence cannot yet be estimated. This axis would measure citation reach, downstream uptake, and recognition once the indexed record accumulates.
Focus0/100
Focus is scored 0, reflecting a deliberately spread indexed profile: the OpenAlex field distribution splits exactly 50/50 between neural operators for parametric PDEs and computational fluid dynamics, and five topic clusters appear — led by Model Reduction and Neural Networks (3), with single counts in CFD and aerodynamics, neural network applications, meteorological phenomena and simulations, and multi-objective optimization. The cataloged papers themselves are tightly scoped, but the indexed footprint is diverse.
Collaboration53/100
The collaboration score of 53 rests on 13 distinct coauthors in the recent window and a recurring circle of 12 named collaborators. Both cataloged 2026 papers are multi-authored, with Pan in non-corresponding positions (last author on the pretraining study, middle author on the operator-recast paper), suggesting contribution within team projects. The coauthor count is high relative to only 3 recent works, indicating large per-paper teams.
Activity40/100
Activity scores 40: OpenAlex records 3 works in the last three years, all dated 2026, with no earlier year-level counts in the snapshot. The visible output consists of recent arXiv preprints, so the record is too short to establish a sustained cadence; the 2026 cluster points to an active current phase rather than a long publication history.
Reproducibility78/100
Reproducibility is the strongest measured axis at 78, the mean across the 2 papers Evolve assessed. The operator-recast paper rates high reproducibility on the strength of broad benchmarking, while the pretraining and distribution-shift study rates medium despite careful statistics, being limited to one architecture and dataset. Both arXiv preprints report explicit benchmark suites and stated limitations, consistent with an open, testable working style.
Research summary
Shaowu Pan is a researcher at Rensselaer Polytechnic Institute (US) working in scientific machine learning, with an OpenAlex field profile split evenly (0.5/0.5) between neural operators for parametric PDEs and computational fluid dynamics. The cataloged work targets reliability of neural PDE surrogates. 'Equation Recast for Canonical Operator Learning Across Parametric PDEs' (2026, middle author) rewrites a family of parametric PDEs as one canonical operator plus an analytic fixed-point correction, is benchmarked broadly, and probes zero-shot extrapolation. 'How Does Distribution Shift Shape Pretraining Gains in Neural PDE Surrogates?' (2026, last author) shows pretraining gains hinge on shift type and target data budget in RANS turbulence-closure surrogates. OpenAlex topic clusters — Model Reduction and Neural Networks (3), CFD and Aerodynamics, Neural Networks and Applications, Meteorological Phenomena and Simulations, and Multi-Objective Optimization — outline a simulation-driven agenda. The indexed record is early-stage: 3 works in the last three years (all dated 2026), 13 distinct coauthors, and h-index and i10-index of 0, so output and impact are unscored. Reproducibility is the strongest measured axis, a 78/100 mean across 2 Evolve-assessed papers.
Collaboration pattern
Pan works within multi-author teams: both 2026 catalog papers are co-authored, with Pan in non-corresponding positions — last author on the distribution-shift study and middle author on the operator-recast paper. Twelve recurring collaborators are listed, including Qiyun Cheng, Valentin Duruisseaux, Cesar F. Clauser, and Anima Anandkumar, spanning optimization and simulation-related areas. Thirteen distinct coauthors appear in the recent OpenAlex window against only 3 recent works, pointing to per-paper teams larger than the output volume.
Research keywords
Highlights
Recasting parametric PDE learning as one canonical operator
When pretraining actually pays for neural PDE surrogates
Timeline
- 2026Co-authored 'Equation Recast for Canonical Operator Learning Across Parametric PDEs' (arXiv, middle author), recasting parametric PDE learning as one canonical operator with an analytic fixed-point correction.
- 2026Co-authored 'How Does Distribution Shift Shape Pretraining Gains in Neural PDE Surrogates?' (arXiv) as last author, quantifying pretraining gains under distribution shift for RANS turbulence-closure surrogates.
- 2026OpenAlex logs 3 works in 2026 with 13 distinct coauthors, marking the visible start of the indexed record; h-index and i10-index both stand at 0.
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 are from OpenAlex only; the profile is name-resolved from public bibliographic data, and no ORCID is linked.
- Citation aggregates (h-index 0, i10-index 0) reflect a sparse snapshot, so output and impact axes are unscored rather than zero-valued.
- The reviewed catalog covers only 2 arXiv preprints from 2026; the reproducibility mean of 78 rests on just 2 assessed papers.
- Co-author affiliations, roles, and funding context are not in the input data; collaboration descriptions rely on author positions in the cataloged papers.