Kevin Mancini

MAP

01

Alma Mater Studiorum – University of Bologna

Bologna, IT (2018-2021)

Bachelor’s degree in Computer Science
02

University of Hull

Hull, UK (2021-2022)

Exchange Program
03

Deloitte

Milan, IT (2022-2023)

Consulting
04

Imperial College London

London, UK (2023-2024)

Master’s studies in Computer Science
05

Sapienza University of Rome

Rome, IT (2024-2025)

Master’s studies in Applied Mathematics
06

Technical University of Munich

Munich, DE (2025-2026)

Exchange Program
07

University of Warwick

Coventry, UK (2026-2030)

PhD in Mathematics

PUBLICATIONS

Kevin Mancini, Islem Rekik · arXiv preprint · 2025

DeltaGNN: Graph Neural Network with Information Flow Control

Introduces information flow control and the information flow score to mitigate over-smoothing and over-squashing in graph neural networks while keeping the added computational overhead linear.

arXiv

Kevin Mancini, Islem Rekik · Graphs in Biomedical Image Analysis (MICCAI GRAIL) · 2025

DuoGNN: Topology-aware Graph Neural Network with Homophily and Heterophily Interaction-Decoupling

A topology-aware architecture that separates homophilic and heterophilic interactions so they can be modeled with different inductive biases across heterogeneous graph benchmarks.

arXiv YouTube Journal

Kevin Mancini, Mirko Viroli, Gianluca Aguzzi · Bachelor thesis, University of Bologna · 2022

ScaFi: Integration and Performance Analysis with Scala Native

Bachelor thesis on integrating ScaFi with Scala Native and evaluating performance for aggregate computing workflows in distributed systems.

Journal