Emiliano Pallotti

Fondazione "Ugo Bordoni"

Unclaimed · AI-structured profileOpenAlex

Learning-to-optimize researcher applying GNNs and MILP to AI-native 6G, NR-V2X relaying and FM spectrum coordination

papers here

reviewed by Evolve

first-author

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—

No publication-volume data is available in this snapshot, so output cannot be scored. This axis would measure the number and cadence of catalogued publications; the three 2026 arXiv preprints listed here are a catalog sample, not a verified total.

Impact—

No citation or uptake data is available, so impact is not scored. This axis would measure how the work influences other researchers — citations, references and downstream use; nothing in the current snapshot evidences the reception of these three preprints.

Focus0/100

The focus axis scores 0 because the field distribution is perfectly flat: the three catalogued papers divide exactly into thirds — 0.33 each for AI-native 6G network architecture / network optimization, wireless networking / AI for communications, and wireless communications / radio spectrum management. This is breadth without a dominant core in the current snapshot: one architectural position paper, one AI-methods paper and one spectrum-management paper. A larger sample would show whether one direction consolidates.

Collaboration32/100

The collaboration score of 32 rests on a very small raw base: four distinct coauthors across the three catalogued papers, forming a compact, stable team around Federica Mangiatordi, Giambattista Amati, Simone Angelini and Pierpaolo Salvo. There is no visible external network yet. His authorship positions — middle on two papers and last on one, with no corresponding-author credit — suggest a contributing member of an in-house group rather than an independent hub.

Activity—

No productivity time series is available, so activity cannot be scored. This axis would measure publishing cadence and recent engagement over time; all three catalogued papers carry 2026 dates, but no year-by-year record exists to compute a trend.

Reproducibility32/100

Reproducibility is the best-evidenced axis: a mean of 32/100 across three scored papers (reproPapers 3), with per-paper readiness of low, low and medium. Evolve's verdicts explain the ceiling — the architecture paper imports its results from prior work, the FM study rests on one confidential scenario, and the NR-V2X relay paper, the strongest of the three, lacks baselines and error bars. Released code, data and baselines would move this score fastest.

Research summary

Emiliano Pallotti is a researcher at Fondazione "Ugo Bordoni" whose catalogued work sits at the intersection of optimisation methodology and wireless network engineering. The Evolve catalog lists three 2026 arXiv preprints, and his footprint is evenly split — exactly one third each — across AI-native 6G network architecture / network optimization, wireless networking / AI for communications, and wireless communications / radio spectrum management. The unifying thread is Learning-to-Optimize: trained surrogate models that stand in for exact optimisation. An architectural position paper argues that L2O is the missing architectural layer of AI-native networks, illustrated with an NR-V2X GNN case. A companion study trains a MILP-supervised GINE graph neural network for real-time NR-V2X relay selection, claiming near-optimal decisions with roughly a 100x speed-up, though without baselines or error bars. The FM spectrum paper splits cross-border power control over Top-K interference graphs, with clustered simulated annealing matching full simulated annealing at about 6% of the runtime — evidence drawn from a single confidential scenario. He operates inside a compact team of four distinct coauthors — Federica Mangiatordi, Giambattista Amati, Simone Angelini and Pierpaolo Salvo — in middle or last author positions, so far without corresponding-author credit. Reproducibility readiness is modest, averaging 32/100 across the three papers. Output, impact and activity metrics are not yet available in this snapshot.

Collaboration pattern

Collaboration is concentrated in a single four-person team: Federica Mangiatordi, Giambattista Amati, Simone Angelini and Pierpaolo Salvo account for the distinct-coauthor count of 4 across the three papers. Pallotti appears in middle position on two papers and last on one, never as corresponding author — consistent with a contributing specialist inside Fondazione Ugo Bordoni's in-house group rather than an externally networked PI. No cross-institution partners are visible in this snapshot.

Research keywords

Learning-to-OptimizeAI-native networks6G architecturesurrogate modelsGraph neural networksMILPNR-V2X relay selectionFM broadcastingSpectrum coordinationCross-border interferenceInterference graphsWireless network optimization

Highlights

Learning-to-Optimize framed as the missing architectural layer of AI-native networks
The position paper argues L2O should be treated as a first-class architectural layer of AI-native 6G networks, demonstrated through an NR-V2X GNN case; Evolve notes the results are imported from the authors' prior papers, and reproducibility readiness is rated low.
Clustered simulated annealing matches full SA on FM power control at ~6% runtime
For FM cross-border coordination, Top-K interference graphs decompose the power-control problem; clustered simulated annealing matches the full SA solution at about 6% of the runtime, though the evidence rests on one confidential scenario and readiness is rated low.
MILP-supervised GINE claims near-optimal real-time NR-V2X relay selection
A GINE graph neural network supervised by MILP solutions delivers claimed near-optimal relay selection with a ~100x speed-up; Evolve calls it a credible simulation study, but notes the absence of baselines and error bars (readiness: medium).

Timeline

2026Published an AI-driven study of real-time relay selection in smart urban NR-V2X networks via learning-to-optimise graph neural networks (MILP-supervised GINE, ~100x speed-up claim).
2026Published interference-driven clustered optimisation for FM spectrum coordination: Top-K interference graphs with clustered simulated annealing matching full SA at ~6% runtime.
2026Published the architectural position paper proposing Learning-to-Optimize as the missing architectural layer of AI-native networks, illustrated with an NR-V2X GNN case.

Papers in the Evolve catalog

Frequent co-authors

Federica Mangiatordi×3 Giambattista Amati×2 Simone Angelini×2 Pierpaolo Salvo×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.

Radar scores are computed from a bibliometric snapshot; the output, impact and activity axes lack underlying data and are not scored.
All three catalogued papers are 2026 arXiv preprints; no peer-reviewed venue records are in scope yet.
No ORCID or country data was provided; the profile is name-resolved from public bibliographic data, and "Emiliano Pallotti" may be a shared name.
The reproducibility score is a mean of 32/100 over three papers; per-paper readiness is low, low and medium, per Evolve review verdicts.