Every project has a story, and ENCODING’s began long before the proposal was submitted.
In this interview, project coordinator Alessandro Parente (ULB) shares how an idea discussed after a conference in Aachen turned into a successful Marie Skłodowska-Curie Doctoral Network.
From the first brainstorming sessions and the origin of the acronym to the lessons learned along the way, Alessandro reflects on the challenges, motivations, and the people who helped shape ENCODING into what it is today.
What motivated you to participate in this Marie Curie project?
Marie Curies ITNs (now DNs) are projects I have also very much appreciated. The focus is on science and on forming brilliant people, which I consider an ideal setting for a researcher.
Who were the “brainstormers” who came up with the idea?
I remember the first submission happened in 2020. We started discussing in 2019 with Heinz, Antonio (at that time at RWTH), Aurélie (at that time in ULB), and Mara. I specifically remember the discussion when we picked the acronym. We had been in KAUST with Aurélie and Antonio, and we were flying back to Brussels. At the airport in Jeddah, during the long 3 hours we waited for the airplane to board, we went through 20 or more acronyms (without ChatGPT) until we landed on ENCODING.
The acronym was perfect because we really felt, already at that time, the need to propose machine learning models that could encode (and not neglect) the complex physics of combustion.
How did the initial idea come about? Was it a flash of inspiration, a meeting, or just coffee and chat?
That year, I had the opportunity to give a keynote at the International Conference on Numerical Combustion in Aachen on data driven combustion techniques I had been investigating for many years, since my doctoral stay in Utah with Phil Smith and James Sutherland. After the conference, I thought that could be a subject worth of an ambitious project and I proposed that to Heinz. After many rounds (Heinz loves to discuss!), we thought we had the good idea: hybrid physics-based, machine learning models. This is quite obvious today, but wasn’t so back then when people was getting very excited with complex blackbox models.
The sketch below is the refined version of the idea, that we sketched with Heinz in Gerard Degrez’s office, where he stayed for a month as holder of the Chair Jaumotte’s chair.
What was the biggest challenge when getting the project started?
For starters, we had to submit twice! I think the first submission was a bit rushed. We were out of time and I think the proposal suffered form a lack of coherence. The time we got for the resubmission allowed us to make it clearer and more coherent, particularly for what concerned the interaction between the teams and the individual doctoral candidates.
Was it easy to find and select our DCs?
In my case, yes, if we set aside the issues related to visa and other administrative procedures associated with the admission of non-EU residents. This made me reflect more broadly about the luck we have in moving freely (which is often underestimated by the populist lovers) and, in contrast, the hassle many people from the world encounter to pursue their dreams and interests.
Any memorable or funny moments with the DCs you would like to share?
All of them, every moment is quite memorable, I just do not remember them! 🙂 There is maybe one, though, during the kick-off in Brussels, where we also had a school for the doctoral candidates. But I will not comment on it!
If you could go back, would you do anything differently in the selection or mentorship of the DCs?
We always think we could have done things better or in a different way. But, honestly, I cannot complain about anything major. All the DCs I supervise are hard-working and nice people. I think they think the same about me.
What has surprised you the most about how the project has evolved so far?
I am the coordinator, and I am very pleased with how the different teams work together. We are all very busy PIs, and nevertheless, all the important milestones are fulfilled, showing respect for each other and a common goal.
What would you improve in the management or coordination if you could do it again?
Nothing, our PM has been and is great!
Any lessons learned you’d like to share with other PIs or future projects?
The sole hiccup we had up to now was related to the drop of a DC, which is apparently a common issue with MCs. Not sure if any measure would have prevented that, but I would suggest that all PIs have their DC highly integrated within their local work environment. It is important that any DC can discuss and exchange with other peers daily.
What do you hope to see in the next phase of the project?
We have reached important results in experiments and simulations, as promised in the original proposal. I am very proud of this. I look forward to seeing all the DCs graduate!
If you had a superpower to improve the project, what would it be?
Telepathy! I would sometimes get into people’s minds and understand what they (really) think. For all the rest, we are well equipped.
If you had to describe this project with a movie, song, or character, what would it be?
It took me time to think, and it is quite difficult to find a correspondence! Now, since I recently watched that movie (again.. I watched it 5 times), I will say Interstellar. This is a long shot since we have no black holes in ENCODING, but I like the idea that no model or machine can replace the human being. Our DCs need to be the future Murph(s)!
What is your “secret ritual” to stay motivated during long projects?
That is a difficult one. I think I get most of my motivation from the students. They are young, curious, and come up with great ideas (not all the time, hum!), and this is great fuel.
