“I learned that the bedtime rules didn’t count when we were soldering”, Sigrid Passano Hellan laughs, recalling memories from her childhood with her father.
“My father is a radio amateur and had some small electronics at home. It could be a small radio that he kept in a mint tin. And then we would sit there with the soldering iron. It was both exciting to be involved, but also fun to be allowed to stay up later than usual, she says.
Work in progress
Hellan started as a researcher at NORCE in August, where she works on how machine learning and artificial intelligence can be used in climate research.
“It is very exciting to work with the combination of weather and machine learning. Large machine learning models have been developed for weather prediction by major technology companies. Now, meteorological institutes are trying to use these models instead of numerical and physical models, so there is a significant focus on the field”.
Hellan’s interest in machine learning came during her master’s in London, where she studied electrical engineering at Imperial College.
“There was quite a wide range in the program, so we got to work with both soldering and programming at different levels, from machine codes to machine learning,” says Hellan.
Trendy
Hellan adds that machine learning was very trendy at the time. She wrote her master’s thesis on optimizing the power system. After that, Hellan wanted to continue with machine learning and apply it to climate and environmental issues. Therefore, she found two supervisors with an interest in this at the University of Edinburgh.
“There, I looked at machine learning for optimization. How to automatically plan sensor networks for air pollution, in addition to using this method to optimize other machine learning models. It was a lot of fun, and we had the possibility to try what seemed exciting”, says Hellan.
Air quality
A part of her work in Edinburgh was to look at the air pollution in London. After that, she has taken with her what she knows about air pollution.
“That was actually something I checked when I moved to Bergen. How the air pollution was in the city. I don’t want to live somewhere with poor air quality. I think we should all learn more about air quality in general.”
Hellan is having a good time in Bergen. She finds it exciting to be at this stage in AI development. And the development is progressing rapidly.
“It will change how we work and what tools we use. With weather models, it is quite possible that weather forecasting will be generated by AI, but we still need researchers to determine when the models are accurate and when it is not. And most importantly, what it means for people, Hellan says.
Need experts
AI models introduce a new potential issue since they predict the future based on what they’ve seen in past data. This can make it difficult for AI to conceive of future extremes that haven’t occurred in the recent past.
“For today's weather, we have a lot of data on how it usually is. However, we don't have data for the future climate. We can generate data on how we think the climate will be, but we need experts who know the area well and can access whether it seems reasonable and what the assumptions are”, she says.
When she is home in Trondheim, she and her dad sometimes work on things together. They built a 3D printer the year before she started her studies. “The building part is always the most fun. We can sit and solder, maybe make something that blinks, like Christmas tree lights. Now we have reached a point where it is not enough to buy one soldering kit and give it to my dad. We need two kits so that we both can work on it, she laughs.