Multimodal machine learning to help industry stay one step ahead

When did you start your position as an industrial PhD student? 

I started in November. In the beginning, I’ve spent a large part of my time getting familiar with the research I’ll be focusing on during my PhD. I’m employed by RISE and will be working with companies like Valmet. My research is supervised by Professor Jan Lundgren at Mid Sweden University.

Can you tell us a bit about your background?

I’m originally an energy engineer, I got my Bachelor degree at the University of Rome Tor Vergata, and for my Master degree I was part of a Nordic master program which involved studying at two Nordic universities. I spent one year in Reykjavik University in Iceland and completed my final year at Chalmers University of Technology in Gothenburg. That’s where I met my wife. When she got a job in Sundsvall, we moved to Västernorrland.

Here, I’ve worked as a research and development engineer at Agtira (which was then called Peckas Naturodlingar) a company that at my time focused on aquaponics, but now develops tailor-made greenhouse systems using advanced technology to enable stores to offer locally grown vegetables year-round. I’ve also held the same role at Absolicon, a company specializing in solar thermal energy and district heating.

What will be the focus of your research as an industrial PhD student?

Although I have a background in energy engineering, I discovered early on that I have a passion for programming. I’ve taken the initiative to teach myself programming, AI, and machine learning - and that will be the focus of my PhD studies.

My research focuses on something called multimodal machine learning. This involves combining data from different types of sources - such as sound and images - to use machine learning for identifying, for example, when a machine part needs to be replaced. That is one of many possible applications of multimodal machine learning.