Rafael Pina
—
First-author researcher split evenly between physiological-signal latent modeling and multi-agent RL communication
2
papers here
2
reviewed by Evolve
2
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—
This axis would measure career publication volume and productivity, including paper counts and output trajectory over time. The snapshot provides no aggregate output statistics for Rafael Pina, so it is left unscored rather than estimated.
Impact—
Impact would capture citation-based influence, such as citation counts, h-index, and uptake of the work by other researchers. No citation metrics exist in the provided snapshot, so this axis is reported as missing rather than guessed.
Focus0/100
The focus score is 0, and the raw field distribution explains why: the two cataloged first-author papers split exactly 50/50 between machine learning for healthcare / physiological signal processing and multi-agent reinforcement learning. One paper models vital-sign latent representations; the other optimizes inter-agent communication. This even two-field dispersion yields minimal topical concentration, so the score reflects breadth across unrelated subfields rather than a data gap.
Collaboration28/100
The collaboration score is 28, grounded in 3 distinct coauthors in the snapshot: Varuna De Silva, Mindula Illeperuma and Corentin Artaud. Both cataloged papers are first-authored, indicating a lead position within this compact team, and neither paper lists him as corresponding author. A small, stable coauthor pool points to close-knit team work rather than large multi-institution collaboration; the low score follows directly from the limited distinct-collaborator count.
Activity—
Activity would measure recent publication cadence — how actively the researcher is currently publishing. The snapshot lists only two 2026 arXiv preprints without a full publication history, so current activity cannot be scored.
Reproducibility60/100
Reproducibility scores 60, the mean across 2 assessed papers (reproMean 60, reproPapers 2), and the two cases diverge sharply. The MARL communication paper (MARC) is rated high readiness, though Evolve notes missing random-seed and confidence-interval reporting. The vital-signs latent-modeling paper is rated low readiness, with a gameable mean-distance metric weakening its evaluation. The mid-range average therefore blends one replicable setup with one evaluation-fragile one.
Research summary
Rafael Pina is an early-stage researcher whose Evolve catalog holds two first-author arXiv preprints, both dated 2026, split exactly 50/50 between machine learning for healthcare / physiological signal processing and multi-agent reinforcement learning. The first applies a conditional variational autoencoder with a latent classifier to represent clinical conditions in vital signs from healthy individuals, targeting data augmentation for MIMIC-style physiological time series and wearable sensors; Evolve credits interesting ablations but flags a gameable mean-distance metric, with low reproducibility readiness. The second, MARC, adds an observation-prediction regularizer and DCT compression to attention-based multi-agent communication, reporting significant gains despite weak theoretical grounding and no random-seed or confidence-interval reporting; its reproducibility readiness is high. The bibliometric footprint is compact: 3 distinct coauthors (Varuna De Silva, Mindula Illeperuma, Corentin Artaud) yield a collaboration score of 28, the even two-field split yields a focus score of 0, and reproducibility averages 60 across 2 assessed papers. Output, impact and activity axes lack sufficient data. With no institution, country or ORCID attached, this profile is name-resolved from public bibliographic data.
Collaboration pattern
Collaboration is compact: 3 distinct coauthors in the snapshot — Varuna De Silva, Mindula Illeperuma and Corentin Artaud — supporting a collaboration score of 28. Both cataloged papers carry first-authorship without corresponding-author status, consistent with a lead-contributor role inside a small, stable team. No large multi-institution networks appear in the available data, and the coauthor-to-paper mapping is not provided, so the exact pairings behind the 50/50 field split remain unresolved.
Research keywords
Highlights
Latent reshaping of healthy vital signs toward clinical conditions
MARC: robust, compressed communication for multi-agent learning
Timeline
- 2026First cataloged first-author preprint: 'Representing Clinical Conditions on Vital Signs from Healthy Individuals using Latent Modeling' appears on arXiv in the machine learning for healthcare / physiological signal processing field.
- 2026Second first-author preprint, 'Robust and Efficient Communication for Multi-Agent Learning' (MARC), appears on arXiv in multi-agent reinforcement learning, completing the 50/50 two-field split of the catalog.
- 2026Bibliometric snapshot: 2 papers in the Evolve catalog, 3 distinct coauthors, collaboration 28, focus 0, reproducibility mean 60 across 2 assessed papers; output, impact and activity lack sufficient data.
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.
- Profile is name-resolved from public bibliographic data; no institution, country or ORCID is attached in the snapshot.
- Bibliometrics cover only the Evolve-reviewed slice: 2 arXiv preprints from 2026; no peer-reviewed venue, citation or publication-history data is included.
- Output, impact and activity axes are unscored because the underlying aggregate metrics are unavailable; no values are estimated.