Giuseppe is a PhD student from the University of Catania in Italy.
He is on a six-month research visit, from April until October 2025, under the supervision of Professor Jianhua Zhang at the Department of Computer Science at OsloMet. His research stay is funded by an Italian national PhD scholarship.
For his PhD thesis research, he is enrolled in a national PhD programme in Italy for sustainable development and climate change.
La Spina originally comes from Catania in Italy and has a background in electronics and automation engineering. Before he came to OsloMet, he was a visiting researcher at the Italian National Institute of Geophysics and Volcanology (INGV), whose main focus is on the active volcano in Catania, Etna.
Assessing and mitigating risks for volcanic eruption
The group at the INGV with which La Spina was affiliated works on modelling and simulations of geophysical flows in order to assess and mitigate risks of volcanic eruption in the region of Catania. These simulations can, for example, predict where lava flow will deposit itself, which can be a key factor in determining where to build residential areas.
“This simulation is computationally demanding and requires a lot of computing resources. Sometimes you have very accurate simulations, but only after the lava flow has already happened. It’s beautiful, but it’s not useful”, he explains.
Avoiding the loss of people and properties
For his thesis, La Spina is exploring the possibility of using machine learning models to identify and simulate complex and large-scale dynamical systems, which account for the generation of geophysical flows.
“We had the idea that artificial intelligence techniques could be leveraged for improving the simulations for volcanic activity monitoring”, he explains.
During his studies, he met professor Jianhua Zhang, who is an expert in the field of artificial intelligence at OsloMet and was a visiting professor at the University of Catania in the autumn of 2024.
Exploring the potential of machine learning
“We want to move from a physics-based approach to a data-driven machine learning approach, so that we improve the accuracy and the efficiency of the simulations. We’re exploring the potential of machine learning to solve this large-scale and complex system modelling and simulation problem”, says Zhang.
“In that way, we can predict volcanic eruptions and avoid the loss of people and properties due to natural disaster”, he adds.
Zhang thinks that most researchers in the field are still focused on computer-based simulations instead of the development and deployment of real monitoring systems for the forecasting and prevention of natural disasters.
“But the development of such systems must be based on solid results from a lot of preliminary research work, and that’s what this PhD thesis is centred on”, he adds.
Better results in working together internationally
“Geophysical flows are intrinsically a multidisciplinary topic of research, because physics, chemistry and geophysics are involved in studying how a volcano works, and how to predict catastrophic events”, La Spina explains.
He came to Oslo in April 2025, which also happens to be his first time in Scandinavia. La Spina sees a lot of value in research collaboration across nations.
“Working together means achieving better results, faster. We merge our knowledge and skills together”, he explains.
Zhang visited the University of Catania and the Italian National Institute for Geophysics and Volcanology last year. He thinks that the university specifically values international cooperations.
“I’m very impressed by the research environment and culture in Italy. Internationalisation is one of OsloMet’s strategic priorities, and we always want to foster and strengthen our international research partnerships, specifically through such substantial research collaboration”, he explains.
He thinks that researchers in geophysics have already realised and started to appreciate the power of artificial intelligence (AI) in solving their long-standing problems.
“AI can bring new ideas, methods and a much-improved performance for the modelling, analysis and simulation of complex dynamical systems. We’re trying to strike a balance between the accuracy, computational tractability and interpretability of the machine learning model-based solutions”, he says.
The photo at the top of the article shows OsloMet Professor Jianhua Zhang to the left, and Italian PhD student Guiseppe La Spina.
