Robert Jenssen

Centre for Arctic Gas Hydrate, Environment and Climate · Denmark

Unclaimed · AI-structured profileORCID OpenAlex

ML researcher spanning multimodal medical AI and computer vision, with deep roots in clustering and Bayesian methods

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

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—

No output score is available yet. This axis would measure publication volume and productivity trends — total works, annual counts, and author-role distribution — from the bibliometric record; it is pending computation and is intentionally not estimated here.

Impact—

No impact score is available yet. This axis would measure citation-based influence — citations per work, citation trajectories, and uptake over time — and is pending computation; it is not estimated from the h-index reported here.

Focus0/100

The focus axis scores 0, reflecting an unusually distributed topical footprint rather than a concentrated core. The OpenAlex topic record spreads across advanced clustering algorithms (49 works), neural networks and applications (47), advanced neural network applications (37), face and expression recognition (35), and Bayesian methods and mixture models (31), while the recent field split is exactly even at 0.5 machine learning for medical diagnosis and 0.5 computer vision. No single topic dominates the record.

Collaboration54/100

Collaboration scores 54. The recent three-year window contains 53 works with 14 distinct co-authors, indicating a moderately sized, recurring network rather than large-scale multi-team consortia. Frequent collaborators in the record include Michael Kampffmeyer, Shujian Yu, Changkyu Choi, Cheng Zhang, Fan Bai, Zhaozhao Ma, and Yuqing Yang. Neither of the two Evolve-reviewed 2026 papers lists Jenssen as corresponding author, one in a middle position and one in the last position.

Activity100/100

Activity scores the maximum 100. Indexed output runs continuously from 2001 through 2026, with 53 works in the last three years and a strong recent climb: 13 works in 2023, 11 in 2024, 23 in 2025, and 19 already in 2026 — a partial year. The record's single annual peak remains 32 works in 2017, but the 2023–2026 stretch shows sustained, high-frequency publishing.

Reproducibility55/100

Reproducibility scores 55, computed from two Evolve-assessed papers (reproMean 55 across reproPapers 2). Both 2026 arXiv preprints carry a medium repro-readiness rating. Evolve's review notes that SMILE leads headline metrics on five datasets but not every AUC, with some internal inconsistencies, while the video-QA benchmark rests on small datasets — notably a 20-video CMD split — that limit statistical strength.

Research summary

Robert Jenssen is profiled here from public bibliographic data (OpenAlex identity, ORCID 0000-0002-7496-8474), associated in the source record with the Centre for Arctic Gas Hydrate, Environment and Climate. The record shows a long-running machine learning career: h-index 40, i10-index 103, and indexed works in every year from 2001 to 2026. Topic modelling of the record concentrates on advanced clustering algorithms (49 works), neural networks and applications (47), advanced neural network applications (37), face and expression recognition (35), and Bayesian methods and mixture models (31). The recent record splits evenly between two fields — machine learning for medical diagnosis (multimodal medical AI) and computer vision (multimodal machine learning) at 0.5 each — and the last three years contribute 53 works with 14 distinct co-authors. Two 2026 arXiv preprints reviewed by Evolve sit in this space: SMILE, an information-bottleneck-based self-explainable multimodal diagnosis framework that leads headline metrics on five datasets, and a visual-question-answering benchmark showing that frontier video models such as Gemini collapse from 74% to 30% when multiple-choice options become visual frames. On the Evolve radar, activity scores 100, reproducibility 55, and collaboration 54, while focus scores 0 given the deliberately broad topical spread; output and impact scores are not yet computed.

Collaboration pattern

Collaboration scores 54: 53 works over the last three years with 14 distinct co-authors suggest a stable, moderately sized network rather than broad multi-team consortia. Recurring names include Michael Kampffmeyer, Shujian Yu, Changkyu Choi, Cheng Zhang, Fan Bai, Zhaozhao Ma, Yuqing Yang, Alexander Schmatz, and Rwiddhi Chakraborty. In the two Evolve-reviewed 2026 papers Jenssen is a middle author on SMILE and last author on the video-QA benchmark, neither as corresponding author.

Research keywords

clustering algorithmsneural networks and applicationsmultimodal machine learningmedical diagnosis AIcomputer visionBayesian methods and mixture modelsinformation bottleneckself-explainable AIvisual question answeringvideo-language modelsface and expression recognitionbenchmark evaluation

Highlights

SMILE: information-bottleneck-based self-explainable multimodal medical diagnosis
This 2026 arXiv preprint proposes an information-bottleneck framework for self-explainable multimodal diagnosis. Evolve's review finds it leads headline metrics across five datasets — though not every AUC — and flags some internal inconsistencies; repro-readiness is rated medium.
Benchmark shows frontier video models collapse when answers become visual
In this 2026 arXiv benchmark paper (Jenssen in the last author position), Gemini drops from 74% to 30% when multiple-choice options become visual frames rather than text. Evolve calls the design compelling but notes that small datasets — including a 20-video CMD split — limit statistics.
Two decades of machine learning output with an h-index of 40
The OpenAlex record spans 2001–2026 without a gap, and 103 works have at least 10 citations (i10-index). Topic counts place clustering algorithms (49) and neural networks (47, plus 37 in advanced applications) at the core of the profile.

Timeline

2001Earliest year with indexed works in the OpenAlex counts-by-year record (1 work).
2006An early output peak with 13 indexed works in a single year.
2017Highest annual count in the record: 32 indexed works.
202523 indexed works — the strongest year since the 2017 peak.
202619 indexed works so far this year; two Evolve-reviewed arXiv preprints — SMILE and a visual-question-answering benchmark for video models — enter the catalog.

Papers in the Evolve catalog

Frequent co-authors

Yuqing Yang×1 Alexander Schmatz×1 Zhaozhao Ma×1 Changkyu Choi×1 Shujian Yu×1 Rwiddhi Chakraborty (Oliver)×1 Yinong (Oliver)×1 Wang×1 Cheng Zhang×1 Fan Bai×1 Zhuoran You×1 Michael Kampffmeyer×1

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 (h-index, i10-index, counts by year, topics) come from an OpenAlex record; output and impact radar scores are not yet computed.
The profile is name-resolved from public bibliographic data; the affiliation is as given in the input, and record–institution alignment was not independently verified.
Reproducibility (55) and the highlights rest on only two Evolve-reviewed papers, both 2026 arXiv preprints — a small and very recent sample.
The 2026 annual count (19) covers a partial year and is not comparable with completed years.