Neural-operator and foundation-model surrogates for wave and hydrologic dynamics at NRL Ocean Sciences
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-15) 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—
No output metrics are available yet: publication counts and per-year volume are not in the current snapshot, so this axis, which would measure total and recent research productivity, is left unscored rather than estimated.
Impact—
Impact data such as citations or uptake of the two surrogate models is absent from this snapshot; this axis would measure scholarly influence and real-world adoption of the work, so it remains unscored instead of estimated.
Focus0/100
The focus score is 0, driven by an exactly even field distribution: the snapshot splits the reviewed work 50/50 between scientific machine learning for neural-operator fluid surrogates and scientific machine learning for hydrologic surrogate modeling. Both strands serve the same broader goal of replacing expensive physics solvers, but with only two reviewed papers, the profile reads as balanced across two sub-directions rather than concentrated in one, pending more publications.
Collaboration36/100
The collaboration score is 36, reflecting a compact working circle: the snapshot records only 5 distinct coauthors across the reviewed papers, with Mahdi Abdelguerfi, Enrique Hernandez Noguera, Md Meftahul Ferdaus, Nathan Cooper, and Edward Holmberg recurring. On both 2026 papers Ioup sits in middle-author positions and is not the corresponding author, so the pattern is intra-team contribution within a small NRL-led group rather than an extended external network.
Activity—
Activity cannot be assessed: the snapshot holds only two 2026 papers and no year-by-year publication record, so recent research tempo and continuity, which this axis measures, are left unscored.
Reproducibility78/100
Reproducibility is the best-supported dimension, scoring 78 as the mean across 2 reviewed papers. The DU-NO wave-modeling study is rated high readiness, consistent with capacity-matched controls against U-FNO and a 14.9% gain on its own benchmark, while the KG-Chronos-2 HEC-RAS surrogate is rated medium, its 14.13% RMSE reduction resting on two events from a single project. The overall picture is disciplined, well-controlled methodology with room to broaden evaluation coverage.
Research summary
Elias Ioup is a researcher in the Ocean Sciences Division of the U.S. Naval Research Laboratory working at the intersection of scientific machine learning and geophysical flow modeling. Evolve's catalog contains two of his 2026 arXiv papers, in both of which he appears as a middle author rather than corresponding author, indicating contributions within a team-led effort. His reviewed work splits evenly between two strands of surrogate modeling: neural-operator surrogates for fluid dynamics and hydrologic surrogate modeling. The first strand is represented by DU-NO, a 3.6M-parameter double U-shaped neural operator that Evolve found beat U-FNO by 14.9% on its own wave benchmark under capacity-matched controls, though on single-site data. The second is Knowledge-Graph-Augmented Chronos-2, which adds graph retrieval and gated correction to a frozen Chronos-2 backbone and cut event-balanced RMSE by 14.13% for HEC-RAS surrogate forecasting, with evidence resting on two events of one project. Reproducibility is the strongest measured dimension, averaging 78 across 2 reviewed papers (one high, one medium readiness), while collaboration scores 36 with only 5 distinct coauthors and focus scores 0 on an even 50/50 field split; output, impact, and activity remain unscored for insufficient data.
Collaboration pattern
Ioup works within a compact NRL Ocean Sciences Division team: only 5 distinct coauthors appear in the snapshot, with Mahdi Abdelguerfi, Enrique Hernandez Noguera, Md Meftahul Ferdaus, Nathan Cooper, and Edward Holmberg as recurring collaborators. On both 2026 arXiv papers he is a middle author and not corresponding, suggesting he contributes specialist components such as operator design and retrieval augmentation inside group-led projects rather than leading papers himself.
Research keywords
Highlights
Parameter-efficient neural operator for phase-resolving wave modeling
Knowledge-graph retrieval augments frozen Chronos-2 for HEC-RAS forecasting
Timeline
- 2026Co-authored DU-NO, a parameter-efficient double U-shaped neural operator for phase-resolving wave modeling, released on arXiv.
- 2026Co-authored Knowledge-Graph-Augmented Chronos-2 for HEC-RAS surrogate forecasting, released on arXiv.
- 2026Both papers entered the Evolve review catalog in the same year, with reproducibility readiness rated high and medium respectively, establishing the current two-field snapshot.
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.
- All scores derive from a small bibliometric snapshot of only 2 reviewed papers; output, impact, and activity axes are unscored due to insufficient data.
- Both cataloged papers are 2026 arXiv preprints; peer-reviewed publication status is not recorded in the data.
- No citation counts, h-index, or full publication list were provided, so none are quoted or estimated in this profile.
- Profile is name-resolved from public bibliographic data (ORCID 0009-0002-3699-6762); country metadata is absent.