New doctoral student strengthens research in industrial AI
With a background in embedded systems and industrial IoT, Thomas Bénéteau wants to contribute to robust AI models for industry. At Mid Sweden University, he is now researching how AI can strengthen railway infrastructure and improve operational reliability.
From the Atlantic coast of France to Mid Sweden University, Thomas Bénéteau's path to research has not been straightforward.
Before training as an engineer, he had a background in sound engineering. Later, he chose to return to his studies and trained as an engineer at ENSTA Bretagne in France, where he specialized in embedded systems.
During his studies, he worked as an apprentice at an IoT company and came into contact with artificial intelligence in earnest.
- At first, my interest was sparked by the emergence of tools such as Chat GPT, but through courses in machine learning, I quickly realized that AI is much broader than large language models. It is also about specialized models that can be used in sensors, embedded systems and industrial applications, he says.
It was precisely the opportunity to use data from real, physical systems to understand, predict and improve their function that attracted him to the research.
At Mid Sweden University, he participates in the research projects AIRI and Next Industrial AI, where the focus is on predictive modeling using sensor and weather data. A particular focus is on predicting the temperature of train tracks – an important issue for railway infrastructure, especially during the summer when high temperatures can increase the risk of sun curves and affect road safety.
A key research challenge is to develop AI models that are not only accurate, but also robust enough to work in real-world industrial environments where data is often noisy, incomplete, or site-specific.
He sees great potential for AI in industry, but emphasizes the importance of choosing the right tool for the right problem.
- The most useful AI systems will not always be the most general. For many industrial applications, specialized models are often more relevant.
In addition to his research, he is also looking forward to life in Sweden.
- Everyone talks about winter and how tough it can be. I'm actually looking forward to experiencing it myself – and maybe getting the opportunity to ski in the same city where I live. It would be the first time.