Resarch Intern

Sony AI

Avda. del Portal de l'Àngel, 40, Ciutat Vella, 08002 Bacelona, Spain

**CV** (with a full publication list, academic services, and extra resources for talks)

Email •
GitHub / GitHub •
LinkedIn •
ORCID •
arXiv •
Google Scholar •
Web of Science

I am Alessandro Lonardi, a Research Intern at Sony AI in Barcelona, Spain. Before that, I was a PhD student at the Max Planck Institute for Intelligent Systems in Tübingen, Germany. I did my PhD under the supervision of Caterina De Bacco in the Physics for Inference and Optimization group. My PhD was supported by the International Max Planck Research School for Intelligent Systems (IMPRS-IS), which is part of the Cyber Valley initiative. I obtained my Master’s in Mathematical Engineering at the University of Padova in Padua, Italy. There, I also got my Bachelor’s in Physics.

My research focuses on modeling complex systems, with a strong interest in statistical inference, generative models, and optimization methods. I like working at the interface between physics and computer science by developing mathematical models inspired by statistical physics and implementing them in robust, scalable algorithms. I truly enjoy technical work, but I believe that empirical observations should always drive scientific inquiry. The applications that interest me most lie within AI for science, where inferential predictions and mechanistic models can mutually inform each other.

**Some keywords, in no particular order:** Statistical Physics, Complex Systems, Machine Learning, Statistical Inference, Generative Models, Optimal Transport, Belief Propagation, Knowledge Graphs, XAI.

**Jul 1, 2024:**I am joining Sony AI in Barcelona.**May 1, 2024:**A new preprint Cohesive urban bicycle infrastructure design through optimal transport routing in multilayer networks is on the arXiv.**Apr 26, 2024:**I defended my PhD thesis!**Apr 24, 2024:**Our paper Message-Passing on Hypergraphs: Detectability, Phase Transitions, and Higher-Order Information is in Journal of Statistical Mechanics: Theory and Experiment. Here a tl;dr summary of the paper.**Apr 6, 2024:**I gave the talk Bilevel optimization for traffic mitigation in optimal transport networks, at the Research Seminar on Mathematical Optimization of the Mathematical Optimization for Data Science Group at Saarbrücken University, Germany.**Dec 26, 2023:**Our paper Bilevel Optimization for Traffic Mitigation in Optimal Transport Networks is in Physical Review Letters!

**Dec 1, 2023:**A new preprint is online: Message-Passing on Hypergraphs: Detectability, Phase Transitions, and Higher-Order Information. As a symbolic compensation for the emissions generated by our numerical experiments, Nick and I planted a Hyper Mango 🥭.**Oct 9, 2023:**I do not use Twitter anymore. My social media profile is now @alonardi on LinkedIn.**Jun 28, 2023:**A new preprint is online: Bilevel Optimization for Traffic Mitigation in Optimal Transport Networks.**Jul, 2023:**Two talks at Netsci 2023: Infrastructure adaptation and emergence of loops in network routing with time-dependent loads and Bilevel optimization for flow control in optimal transport networks.**Mar 9, 2023:**I gave a talk for the NetPLACE Seminars series.**Feb 3, 2023:**Our work Infrastructure adaptation and emergence of loops in network routing with time-dependent loads is in Physical Review E!**Jan 20, 2023:**Our work Immiscible Color Flows in Optimal Transport Networks for Image Classification is in Frontiers in Physics! As a symbolic compensation for the emissions generated by our numerical experiments, Diego and I planted a cocoa tree 🌿.**May 6, 2022:**Our work Multicommodity routing optimization for engineering networks is in Scientific Reports!**May 4, 2022:**A new preprint is online: Immiscible Color Flows in Optimal Transport Networks for Image Classification.**Dec 21, 2021:**A new preprint is online: Infrastructure adaptation and emergence of loops in network routing with time-dependent loads.**Oct 13, 2021 - Feb 11, 2022:**I am a teaching assistant for the course of Advanced Probabilistic Machine Learning and Applications (2022), at University of Tübingen.**Oct 4, 2021:**Our work Designing optimal networks for multicommodity transport problem is in Physical Review Research!**Jul 14, 2021:**Our work Optimal Transport in Multilayer Networks for Traffic Flow Optimization has just been published.**Apr 19, 2021 - July 31, 2021:**I was a teaching assistant for the course Advanced Probabilistic Machine Learning and Applications (2021), at University of Tübingen.**Feb 12, 2021:**I joined the International Max Planck Research School for Intelligent Systems!

In reverse chronological order, asterisks denote equal contribution.

The last publication I contributed to is:

**Cohesive urban bicycle infrastructure design through optimal transport routing in multilayer networks**

Alessandro Lonardi, Michael Szell, Caterina De Bacco

arXiv • GitHub

**Designing Networks with Adaptation Rules and Optimal Transport**(PhD thesis)

Alessandro Lonardi

University of Tübingen (2024)

**Message-Passing on Hypergraphs: Detectability, Phase Transitions, and Higher-Order Information**

Nicolò Ruggeri*, Alessandro Lonardi*, Caterina De Bacco

Journal of Statistical Mechanics: Theory and Experiment (4), 043403 (2024) • arXiv • GitHub • CO₂ compensation

**Bilevel Optimization for Traffic Mitigation in Optimal Transport Networks**

Alessandro Lonardi, Caterina De Bacco

Physical Review Letters 131, 267401 (2023) • arXiv • GitHub

**Immiscible Color Flows in Optimal Transport Networks for Image Classification**

Alessandro Lonardi*, Diego Baptista*, Caterina De Bacco

Frontiers in Physics 11:1089114 (2023) • arXiv • GitHub • Poster • CO₂ compensation

**Infrastructure adaptation and emergence of loops in network routing with time-dependent loads**

Alessandro Lonardi, Enrico Facca, Mario Putti, Caterina De Bacco

Physical Review E 107, 024302 (2023) • arXiv • GitHub

**Multicommodity routing optimization for engineering networks**

Alessandro Lonardi, Mario Putti, Caterina De Bacco

Scientific Reports 12, 7474 (2022) • arXiv • GitHub

**Optimal Transport in Multilayer Networks for Traffic Flow Optimization**

Abdullahi Adinoyi Ibrahim, Alessandro Lonardi, Caterina De Bacco

Algorithms, 14(7), 189 (2021) • arXiv • GitHub

**Designing optimal networks for multicommodity transport problem**

Alessandro Lonardi, Enrico Facca, Mario Putti, Caterina De Bacco

Physical Review Research 3, 043010 (2021) • arXiv • GitHub

Last updated in July 2024.