Applied ML researcher at FAU Erlangen-Nürnberg bridging medical imaging AI and power-system machine learning benchmarks
4
papers here
4
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
8
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—
Not yet scored. OpenAlex lists 70 works from 2024–2026 plus an h-index of 8 and i10-index of 8, but the normalized output data for this profile is insufficient; this axis would measure publication volume against field baselines.
Impact—
Not yet scored. An h-index of 8 and i10-index of 8 exist in the raw record, but citation-impact normalization (field-weighted citations, percentiles) is unavailable; this axis would measure the influence of the work relative to peers.
Focus0/100
Scores 100 for focus. The topic profile concentrates on power systems fault detection (18 topic hits) with a coherent medical-imaging strand — imaging techniques (15), segmentation (12), radiomics and machine learning (12), AI in cancer detection (11) — while the field distribution anchors to electric power systems (smart grids) / applied machine learning.
Collaboration62/100
Moderate at 48. OpenAlex records 10 distinct coauthors, and the recurring partner list is an Erlangen-centered cluster — Julian Oelhaf, Christian Bergler, Souhardya Chattopadhyay, Antonia Schoening, Jessica Deuschel, Bitan Bhattacharyya, Andreas Maier, Georg Kordowich and Johann Jäger — pointing to stable local teams rather than a broad international network.
Activity100/100
Maximal activity. OpenAlex attributes 70 works to the last three years — 7 in 2024, 24 in 2025 and 39 in 2026 — a steep ramp after single-digit annual counts from 2017 through 2022 (1, 4, 5, 3, 2, 1). Both Evolve-reviewed papers appeared in 2026.
Reproducibility69/100
Scores 55, computed from a single assessed paper (reproMean 55, reproPapers 1), so the axis is provisional rather than established. Directionally the evidence is consistent: EvEMTBench ships an open, versioned benchmark rated high reproducibility readiness, while the Chronos-2 load-forecasting study is rated medium.
Research summary
Siming Bayer is an applied machine-learning researcher at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany, whose OpenAlex record spans two connected strands: medical imaging AI and machine learning for electric power systems. The topic profile concentrates on power systems fault detection (18 topic hits) alongside a medical-imaging cluster — imaging techniques (15), segmentation (12), radiomics with machine learning (12) and AI in cancer detection (11). Recent activity is exceptional relative to the profile's size: 70 works in 2024–2026 (7, 24 and 39 by year), against single-digit annual counts from 2017–2022, while cumulative indicators remain modest (h-index 8, i10-index 8) — the reason output and impact are not yet scored. Evolve has reviewed two 2026 arXiv papers, both with Bayer as last author: EvEMTBench, an open, versioned benchmark for ML in power system protection (4 grids, 24 tasks) rated high for reproducibility readiness though built on fully synthetic data with generic baselines, and a peak-aware short-term load forecasting study across distribution grid aggregation levels in which the time-series foundation model Chronos-2 leads, with gradient-boosting baselines possibly handicapped by fixed seasonal training windows. The overall picture is a highly active, benchmark-oriented applied ML researcher building evaluation infrastructure for grid ML while maintaining an earlier medical-imaging portfolio.
Collaboration pattern
Collaboration is anchored in a stable Erlangen cluster: OpenAlex counts 10 distinct coauthors, and the recurring names — Julian Oelhaf, Christian Bergler, Souhardya Chattopadhyay, Antonia Schoening, Jessica Deuschel, Bitan Bhattacharyya, Andreas Maier, Georg Kordowich and Johann Jäger — suggest ongoing FAU-based teams. On both Evolve-reviewed 2026 papers Bayer holds the last-author position without being corresponding author, a placement often associated with senior or supervisory roles. The radar collaboration score is 48 (moderate).
Research keywords
Highlights
EvEMTBench: an open, versioned ML benchmark for power system protection
Chronos-2 leads peak-aware short-term load forecasting across three grid levels
A dual-domain portfolio: from medical imaging AI to power-system benchmarks
The OpenAlex topic record shows a sustained medical-imaging strand — imaging techniques (15), segmentation (12), radiomics and machine learning (12), AI in cancer detection (11) — alongside the now most frequent topic, power systems fault detection (18), marking an applied-ML researcher working across two domains.
Timeline
- 2017First recorded output in the OpenAlex record: 1 work, the start of a modest early phase.
- 2019Early-phase peak at 5 works, within a steady run of 4 (2018), 5 (2019), 3 (2020), 2 (2021) and 1 (2022).
- 2024Output recovers to 7 works after a 2023 gap in the record.
- 2025Steep rise to 24 works in one year, the beginning of the current surge.
- 202639 works recorded, including both Evolve-reviewed arXiv papers (EvEMTBench and the peak-aware load forecasting study), each with Bayer as last author.
Papers in the Evolve catalog
EvEMTBench: An Open Benchmark for Machine Learning in Power System Protection
Offline Reinforcement Learning for Distribution-Grid Protection
Peak-Aware Short-Term Load Forecasting Across Distribution Grid Aggregation Levels
A Systematic Evaluation of Machine Learning Methods for Fault Detection and Line Identification in Electrical Power Grids
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 derived from OpenAlex only (identity: openalex); citation-based indicators may lag and carry name-resolution uncertainty.
- The output and impact dimensions are unscored (dimMissing); the h-index of 8 and i10-index of 8 are shown as raw values only.
- The reproducibility score (55) rests on a single assessed paper, so it should be read as provisional.
- Only two papers from the Evolve catalog were reviewed for this profile; the broader 70-work recent output has not been individually assessed.