ML researcher bridging GNN evaluation methods, federated clinical data infrastructure, and computational narratology
3
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 metrics are not yet available for this identity: the snapshot lists an h-index and i10-index of 0, so no reliable score is given. This axis would measure total publication volume, venue quality, and the multi-year trajectory of research output.
Impact—
Impact data is insufficient: no citation counts, h-index, or venue-level influence are available in the snapshot, so this axis is left unscored. It would measure citation-based influence, including citing-work volume, field-weighted citation rates, and academic attention over time.
Focus0/100
The focus score of 0 reflects an unusually even spread: the OpenAlex field distribution splits 0.5 to GNN evaluation methodology and 0.5 to medical informatics (federated learning infrastructure), and the topic list adds topic modeling (2 works) plus NLP, graph neural networks, and AI in games (1 each). Reviewed work spans three distinct fields, so concentration in this window is minimal.
Collaboration79/100
Collaboration is this profile's strongest measured axis at 79: the OpenAlex identity records 52 distinct co-authors, and twelve recurring partners appear in the data, including Katrin Rohrbacher, Björn Nieth, Emmanuelle Salin, and Michaela Mahlberg. All three reviewed 2026 papers are middle-authored team outputs rather than solo work, consistent with a group-based, co-author-heavy working style across projects.
Activity40/100
Activity scores 40, a mid-range value consistent with the raw counts: recent3yWorks is 3 and the countsByYear series shows only 2026 with 3 works. All three Evolve-reviewed papers are 2026 arXiv preprints. The short observation window means the activity estimate rests on a single year of data, so this score should be read as an early snapshot rather than a stable trend.
Reproducibility78/100
Reproducibility scores 78, a mean computed across 2 repro-scored papers in Evolve's review set. Two of the three reviewed papers — the GNN stratification study and the LLM worldbuilding study — carry high reproducibility readiness, while the FL-Net federated learning paper is rated medium because its evaluations are simulated and self-scored. Overall, artifacts and evaluation protocols appear comparatively solid.
Research summary
Bjoern M. Eskofier is profiled here from public OpenAlex bibliographic data; no institution, country, or ORCID is attached to this snapshot. The available evidence — three 2026 arXiv preprints reviewed in the Evolve catalog — shows a contributing co-author (middle position on all three, none corresponding) working across three distinct areas: graph machine learning evaluation methodology, medical informatics infrastructure for federated learning, and computational literary studies of AI-generated fiction. The GNN paper proposes homophily-aware stratified k-fold cross-validation, validated on 15 datasets; the FL-Net paper builds a one-stop federated learning framework for multi-center medical data mining; and the NLP study analyzes setting and narrative space in LLM-generated stories with fine-tuned BERT classifiers. Recorded collaboration is extensive — 52 distinct co-authors and twelve recurring partners — and reproducibility is comparatively strong, with a mean repro-readiness score of 78 across two scored papers. OpenAlex topic counts place topic modeling (2) ahead of single appearances in complex network analysis, NLP, graph neural networks, and AI in games, sketching a methodologically broad, evaluation-oriented profile rather than a narrow specialist.
Collaboration pattern
Eskofier's record shows a strongly team-based pattern: 52 distinct co-authors in the OpenAlex snapshot and twelve recurring partners, including Katrin Rohrbacher, Björn Nieth, Emmanuelle Salin, Michaela Mahlberg, Dario Zanca, Heike Leutheuser, and Jonas Mueller. All three reviewed papers are middle-authored group outputs, suggesting the researcher contributes specialist expertise within larger teams rather than leading as first or corresponding author in this catalog window.
Research keywords
Highlights
Homophily-aware stratification for stable GNN evaluation
FL-Net: a one-stop federated learning shop for medical data
How LLMs build fictional worlds
Timeline
- 2026Co-authored 'How LLMs Build Fictional Worlds', a cross-lingual computational narratology study of setting and narrative space in AI-generated storytelling (arXiv).
- 2026Co-authored 'Not All Nodes Are Created Equal', introducing homophily-aware stratification for more stable GNN evaluation across 15 datasets.
- 2026Co-authored the FL-Net paper, a one-stop federated learning framework for multi-center medical data mining.
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 indicators come from an OpenAlex identity snapshot; h-index and i10-index read 0 there, so the output and impact axes are left unscored rather than estimated.
- Institution, country, and ORCID were not provided; the profile is name-resolved from public bibliographic data and has not been independently verified.
- The evidence base is three 2026 arXiv preprints reviewed by Evolve; peer-reviewed venues and citation records fall outside this snapshot.
- The reproducibility score (mean 78) covers 2 of the 3 reviewed papers and is not a verdict on the researcher's entire body of work.