Smart energy? A challenge for Idiap and Artificial Intelligence
Head of this new research group about Energy, Jérôme Kämpf is a man of many talents. Thanks to his experience as a company founder, as a teacher and as a researcher at both EPFL and HEIA of Fribourg, his vision of his domain is particularly broad. He explains us how the group was created and what are the challenges for artificial intelligence in the energy sector.
Why did you join Idiap and accepted to launch this new research group?
I was working with the CREM and they told me about the job opening. As a physicist, I saw the link between the models used for artificial intelligence and models used in physics in the energy field. Statistical tools of those both domains can add value to themselves. Indeed, during my professional experience, I worked on projects dedicated to 3D energy simulations of buildings in urban areas. In such projects there are so many parameters to consider and so many data to process, that artificial intelligence can really provide an asset to be faster and more efficient. Another example of AI use is linked with the calculation of urban comfort indexes. As those indexes are very subjective, statistical methods of artificial intelligence are very useful.
Do you already have projects you would like to work on?
By the end of February, I only work at Idiap at 20%, but I am already building up my group. I received several applications for exchange PhDs and master students. Currently, I am working on a way to improve the visualisation of energy data of a building, so architects can use these tools and realise the impact of their technical solutions on the global energy balance. I am also working on a project in collaboration with the Federal office for the environment to simulate urban heat islands in the city of Fribourg, as well as on an Innosuisse project to develop a smart controlling system for district heating. My other goal is to submit a Swiss National Science foundation (SNSF) project.
What is the development potential for your research group?
My aim is to offer a holistic vision of energy thanks to the tools of artificial intelligence. Today, we have a lot of various data that are not exploited to enhance our understanding of our energy consumption and its optimisation. For example, more and more connect objects can provide information such as temperature and humidity inside housings. These data are impossible to obtain in another way. Another example from the public area is linked to the land registers and mapping. Data are very heterogeneous and cannot be fully exploited to simulate the environment of a building. Being able to cross data from Google Street View, from the land register and a satellite view using machine learning techniques would be very useful to precisely identify trees around a building. Trees location is having an impact on the building energy balance (shadow, wind stopper, etc). So, yes, there is a huge potential for development.
Website: Group Energy Informatics Idiap