Publications

Each paper is listed once and appears on the sites of all the units that signed it.

26 publications

2026

AI Reconstruction of European Weather From the Euro-Atlantic Regimes

A. Camilletti, Gabriele Franch, Elena Tomasi, Marco Cristoforetti
International Journal of Climatology 46(3), e70216, 2026

A non-linear AI model that turns the indices of the Euro-Atlantic weather regimes into monthly temperature and precipitation anomalies over Europe, where so far only linear methods were used: a building block for sub-seasonal to seasonal forecasts.

Deep learning prediction of cold-side quality characteristics from hot-side infrared sensing in glass bottle manufacturing

Mattia Pujatti, Andrea Di Luca, N. Peghini, P. Calanca, Marco Cristoforetti
Engineering Applications of Artificial Intelligence 184, 116323, 2026

A deep-learning model that predicts the cold-side quality of glass bottles, such as wall thickness, from infrared images taken on the hot side right after forming, so that the line gets an early warning instead of waiting for the containers to cool.

2025

AI for industry at FBK

Alessandro Cimatti, Fabio Antonelli, L. Bentivogli, Marco Cristoforetti, A. Micheli, Fabio Poiesi, Fabio Remondino, A. Susi, S. Tonetta
Joint Proceedings of the Thematic Workshops at Ital-IA 2025, Trieste, 2025

An overview of FBK's research on AI for industry, signed by the heads of the 3DOM, DSIP, OpenIoT and TeV units together with colleagues from other FBK centres.

GPTCast: a weather language model for precipitation nowcasting

Gabriele Franch, Elena Tomasi, Rishabh Wanjari, V. Poli, C. Cardinali, P. P. Alberoni, Marco Cristoforetti
Geoscientific Model Development 18(16), 5351–5371, 2025

A generative model that encodes radar precipitation maps as tokens and predicts their evolution like a language model, producing ensembles of nowcasts.

Innovating Space Operations with AI: The AISHGO Project

N. Policella, J. Martinez, N. Salor Moral, S. Kotisis, K. Helmsauer, T. Göttfert, Marco Cristoforetti, Andrea Gobbi, F. Hegwein, G. De Canio, A.-R. Riise, S. Mejri, V. D. Tran, I. Tanco
18th International Conference on Space Operations (SpaceOps 2025), Montreal, paper 49, 2025

Results of the AISHGO project, which brings AI into the mission control of ESA and DLR for incident classification, predictive maintenance, telemetry anomaly detection and long-term satellite health monitoring.

2024

2023

RobDT: AI-enhanced Digital Twin for Space Exploration Robotic Assets

M. Bozzano, R. Bussola, Marco Cristoforetti, S. Goyal, M. Jonáš, K. Kapellos, A. Micheli, D. Soldà, S. Tonetta, C. Tranoris, A. Valentini
The Use of Artificial Intelligence for Space Applications, Studies in Computational Intelligence, 183–198, 2023

The architecture and first results of RobDT, a digital twin for space robotic assets that blends data-driven, physics-based and symbolic models, adapts them at runtime and supports operations with predictions and what-if analysis.

2022

2021

Proton path reconstruction for proton computed tomography using neural networks

T. Ackernley, G. Casse, Marco Cristoforetti
Physics in Medicine & Biology 66(7), 075015, 2021

A deep neural network that reconstructs proton paths in proton computed tomography: as accurate as the most-likely-path method when only multiple scattering occurs, more accurate when nuclear interactions are present, and faster.

2020

How do Twitter users react to TV broadcasts dedicated to vaccines in Italy?

F. Gesualdo, A. D'Ambrosio, E. Agricola, L. Russo, I. Campagna, B. Ferretti, E. Pandolfi, Marco Cristoforetti, A. E. Tozzi, C. Rizzo
European Journal of Public Health 30(3), 481–486, 2020

Sentiment analysis of the tweets that followed two Italian TV programmes on vaccines, one built on scientific evidence and one hosting anti-vaccine guests, as a signal of vaccine confidence.

2019

Evaluating reproducibility of AI algorithms in digital pathology with DAPPER

A. Bizzego, N. Bussola, M. Chierici, V. Maggio, M. Francescatto, L. Cima, Marco Cristoforetti, G. Jurman, C. Furlanello
PLOS Computational Biology 15(3), e1006269, 2019

DAPPER, a validation framework built on a strict data analysis plan, used to measure how reproducible deep-learning models are in digital pathology, here classifying tissue of origin on whole-slide images from GTEx.

2018