Gianluigi Rozza: “Don’t forget to smile and approach academia with a sense of humour about yourself”
Professor Gianluigi Rozza (Italy, 1977) is widely known for his work in Reduced Order Methods (ROM) for computational mechanics. After graduating cum laude in the field of optimal control at Politecnico di Milano in Italy, he started his PhD in mathematical modelling for cardiovascular flows under supervision of Professor Alfio Quarteroni at École Polytechnique Fédérale de Lausanne (EPFL) in Switzerland. During his PhD, he got the opportunity to go on an exchange period at MIT (USA) with Professor Anthony Patera, the researcher who introduced him twice (!) to ROM. Today, Rozza is Full Professor in Numerical Analysis and the SISSA Director’s Delegate for Valorisation of Knowledge, Innovation, Technology Transfer and Industrial Cooperation.
If you lose the first opportunity, there is a second one
You are clearly involved in many research subjects. What would you describe as your main research focus, and how did you get involved in this topic?
My main research focus is Reduced Order Methods for computational mechanics and it is a funny story how I got there. During my master’s, I was a Socrates student (prior to Erasmus), working on optimal control for PDE’s at EPFL. After obtaining my master’s degree I continued here with a PhD position on the same topic. During my PhD, I got the opportunity to spend some of my time at MIT to work with professor Antony Patera on optimal control and domain decomposition. When I arrived at MIT, professor Patera asked me: “Why don’t you work on Reduced Order Methods?” At that moment, I didn’t remember that this professor had already asked my supervisor the same question during my master thesis. During my master I couldn’t see the relation between optimal control and ROM and I was bound to the time constraint of the thesis. But during my PhD, me and my supervisor thought it would be a good investment. And now you see: It became my research life! I have visited MIT every year during my PhD and after 17 years I’m still working together with professor Patera. During my PhD I could also connect optimal control and ROM, which is now the main focus of my group. So that is the lesson: Sometimes if you lose the first opportunity, there is a second one.
Trieste represents for me the capability of building something at higher demand
After several research positions at both MIT and EPFL, you got the opportunity to start as an assistant professor at SISSA in 2012. How did you handle this change?
After five years as a research assistant at EFPL, I had two opportunities to start a tenure track position, of which one of them was at SISSA. Although the other opportunity might have been better, I was fascinated by the project of the creation of the new applied math division. At first, I was quite scared because I had to do so many tasks: teach, advise students, find cooperation with other partners, create international cooperation, go to industry, etcetera. So in the first year I was afraid and was hiding under my desk, not leaving my office and figuring out what to do. It was a bit of a nightmare. At one point, the school director came and I realized I wasn’t given a good impression: I was so split in things, that I wasn’t aware that I was not explaining myself, I was overloading myself. After this, I realized how I had to change and I started cooperating with industry and creating awareness within the institute. At a certain point I received the ERC grant and then it rolled the right direction. Trieste represents for me the capability of building something at higher demand. I was not really contemplating doing so many activities outside my group, but now I’m involved in a lot.
First of all, congratulations on the European Research Council (ERC) grant for your excellent work in this field. What has this grant meant for you and your group?
The grant was not only a recognition for the work that we achieved on this topic, but merely an opportunity to start building a group around it. With this grant I was able to appoint two assistant professors, several postdocs and PhD students, and a couple of masters projects every year.
When I arrived at SISSA in 2010, the school was mainly known for its excellent work in the field of mathematical analysis. I was one of the first people building a new laboratory of applied mathematics focusing on HPC, data science, numerical analysis, and mathematical modeling. With this grant, I was able to show the importance of what we were doing in this new laboratory.
Science and industry advance with mathematics
Your commitment to showcase the importance of mathematics lead you to be appointed as the Director’s Delegate for Valorisation of Knowledge in 2016. How would you explain the value of knowledge utilization?
During my time at SISSA I noticed there is a huge need for science and not only for technology transfer. I noticed that companies had the idea that institutes like SISSA - a small University focusing mainly on Neuroscience, Physics, and Mathematics - are mainly focusing on basic science that are far from a citizen’s point of view. Therefore, I started reaching out to industry to get in closer contact with companies and created this message: “Science and industry advance with mathematics.” Sometimes, Research & Development institutes are so focused on product and process optimization that they don’t think on long term perspective. That is where institutes like SISSA can be of great value. I devoted lots of my time on it and it was definitely an investment we did with the school, but it worked out well. Companies are actively interacting with us and we have created several best practices which has led to more European grants and opportunities for the younger generation.
It seems like you are very much in place at your position in Trieste. Were you always interested in an academic career?
Reflecting on my life as a student, I considered myself a disaster (laughs). I had no idea what an assistant or associate professor was, or had any idea of tenure tracks. I am not coming from a family with anyone in academia. I also applied to several positions in industry, but there was always a last-minute change why it did not happen. Nothing was planned and I would also advise not to be too focused. Make sure you have an open attitude and be careful with your dreams. Dreams can come true, but it also means that you have to put other passions on hold. If you do decide to stay in academia, don’t forget to smile and approach it with a sense of humour about yourself.
The best way to find new approaches in research is by being in contact with students
You are involved in many topics and working together with different groups. Where do you get your inspiration from?
I am very open, because if I’m not open, I’m lost. I’m always on the lookout for different approaches and people outside my area with an active and lively approach to research that can participate in my projects. One of the best ways to obtain these is by being in contact with students. Which is also something that I enjoy the most of my work. Questions of students are a reservoir of ideas and show different perspectives; not yet polluted by constraints or experience, but only driven by curiosity.
What is your vision on the field of numerical analysis?
We have a clear choice: either focus solely on numerical analysis, which has been one of the most fascinating fields in applied mathematics over the last century, or we look forward. Looking forward means that we need a stronger integration between mathematical and numerical analysis and incorporate the new trends in automatic learning, high-performance computing, and data analysis to create faster solutions. At the same time, we need to keep a strong focus on the physics as we have a strong knowledge on the models and equations. Another pillar is developing better perspectives on inverse and control problems to advance the incorporation of data and mathematical models. This could add a new flavour to numerical analysis by investigating how far the mathematical simulation and collected empirical data relate.
· Gianluigi Rozza will talk about state of the art and perspectives of ROM in computational fluid dynamics next NA seminar October 23rd. Register here for the seminar.