ML researcher building learned distributional critics to expose blind spots in fixed weather-forecast metrics.
1
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-20) 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 and preprint counts for this researcher are insufficient in the current data. This axis would measure the volume and consistency of scholarly output across venues and years, none of which can be scored from the single 2026 entry.
Impact—
Impact data is insufficient for scoring. This axis would measure citation-based influence — citation counts, h-index-style aggregates, and uptake of the learned atmospheric critic by other weather-forecast evaluation work — none of which is available in the current snapshot.
Focus100/100
The focus dimension scores 100, reflecting complete field concentration: every reviewed output indexed in the field distribution falls into a single area, machine learning for weather forecasting (atmospheric science), drawn from one catalog paper. This makes the research agenda easy to read — learned evaluation of forecast realism — but the concentration rests on a very small sample, so the score should be read as directional rather than a long-term pattern.
Collaboration39/100
With a collaboration score of 39, the bibliometric snapshot shows six distinct coauthors, five of whom recur as frequent collaborators: Dmitri Demler, Thierry Meier, Luca Rispoli, Fanny Lehmann, and Joel Oskarsson. The reviewed first-author paper was written within this compact circle, suggesting an early, team-based working pattern rather than a broad international network; larger or solo projects cannot yet be assessed from the available data.
Activity—
Activity cannot be scored yet: the data lacks the recent-publication cadence this axis measures, such as outputs per year or time since the latest preprint. One 2026 entry is too sparse to infer an ongoing research rhythm.
Reproducibility100/100
Reproducibility scores 100, computed as a mean of 100 across one paper with a reviewed high repro-readiness rating. The first-author 2026 preprint uses ERA5 reanalysis as a reference dataset and tests a learned critic against established distributional metrics such as Wasserstein distance, an evaluation setup Evolve rates highly for readiness. With only a single evaluated paper, this perfect score signals promising openness rather than a sustained track record.
Research summary
Younes Elberkennou is an early-stage researcher working at the intersection of machine learning and atmospheric science, with a focus on how weather forecasts should be evaluated. The Evolve catalog currently contains one reviewed output: a first-author 2026 arXiv preprint, 'Every Fixed Metric Has a Blind Spot: A Learned Atmospheric Critic for Scoring Forecast Realism.' The paper argues that any fixed verification metric necessarily overlooks certain failure modes and instead trains a learned critic — built on classifier two-sample tests and f-divergence estimation — that scores the realism of forecasts against ERA5 reanalysis, with distributional metrics such as Wasserstein distance serving as references. Evolve's verdict is that the critic adapts convincingly where fixed metrics fail, while cautioning that the evidence is author-curated and the statistics informal. The paper rates highly on repro-readiness, and the bibliometric snapshot carries a reproducibility score of 100 computed over that single paper. Field concentration is total, with a focus score of 100 resting on one indexed field, machine learning for weather forecasting. The collaboration score of 39 reflects a compact pool of six distinct coauthors, including recurring collaborators Dmitri Demler, Thierry Meier, Luca Rispoli, Fanny Lehmann, and Joel Oskarsson. Output, impact, and activity metrics are not yet available, so this profile is best read as an early snapshot of a researcher whose public record is just beginning to accumulate.
Collaboration pattern
The snapshot shows a small, stable circle: six distinct coauthors, with Dmitri Demler, Thierry Meier, Luca Rispoli, Fanny Lehmann, and Joel Oskarsson recurring as frequent collaborators (the source list repeats Joel Oskarsson). The one reviewed paper is first-authored within this group, consistent with a team-based early-career working pattern. The collaboration score of 39 reflects this limited reach; no cross-institution or senior-collaborator dynamics can be confirmed from the available data.
Research keywords
Highlights
A learned critic for weather-forecast realism
High reproducibility readiness
Highly focused early research agenda
Timeline
- 2026First reviewed output appears: a first-author arXiv preprint proposing a learned atmospheric critic for scoring forecast realism.
- 2026The paper earns a high repro-readiness rating, giving a reproducibility score of 100 in the bibliometric snapshot.
- 2026Field concentration is complete: the single reviewed paper places all indexed output in machine learning for weather forecasting.
- 2026The collaboration snapshot forms around six distinct coauthors, including frequent collaborators Dmitri Demler, Thierry Meier, Luca Rispoli, Fanny Lehmann, and Joel Oskarsson.
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 dimensions are computed from OpenAlex-sourced data only; no institution, country, or ORCID was available for identity verification, so this profile is name-resolved from public bibliographic data.
- The profile rests on a single reviewed paper (2026 arXiv preprint); the output, impact, and activity dimensions lack sufficient data and are not scored.
- The frequent-coauthor list contains one duplicated name (Joel Oskarsson), so the effective number of distinct collaborators may be smaller than recorded.
- The perfect focus and reproducibility scores of 100 each derive from one evaluated paper and should not be read as a sustained track record.