Fabio, can you tell us about your research at LNCC?
First of all, I wanted to say that it’s a pleasure to participate in this interview.
I’ a researcher at LNCC, the National Laboratory of Scientific Computing, which is an institute of the Ministry of Science, Technology and Innovation in Brazil. Its main goal is to develop research in computational modelling, which means to model different phenomena and reproduce them in a computational environment. You have some models that you build and then implement their computational aspects. By doing so, you kind of try to reproduce the phenomenon digitally so that researchers can understand and interpret it.
I coordinate the DEXL Lab at LNCC. We focus on Data Science, Artificial Intelligence and Big Data research, which bring us to work with different scientific problems where those techniques can be applied.
Within DEXL we do have masters and PhD students as we offer graduate courses in computational modeling. I advise master, PhD and even graduate students in their final project. We also have post-doc and some engineers.
We have many projects with industry (Petrobras for instance), and international collaborations. The one with Inria is the most lasting one.
And what are your ties with France?
My first professional link to France was during my PhD. In Brazil we have what we call a “sandwich” PhD program: we do half in our Brazilian institution and half in an institution abroad. I was doing my PhD in PUC Rio, which already had collaboration with Inria. Within that context, I got a scholarship to go to France. I went to be advised by Professor Patrick Valduriez in 1999, and I came back to Brazil in 2001. After that, I went to Switzerland for five years, and we continued to collaborate. In 2009 I came back to Brazil and joined LNCC. We started to have this more formal international collaboration in different projects.
Can you tell us more about the current Associate Team HPDaSc ? Is it a continuation of the previous collaborations between Inria and LNCC? (SciDISC, HOMAR, MUSIC, …)
Exactly. We’ve been developing a continuous collaboration between Inria and LNCC in different projects. The first one may have been HOSCAR. At that time, it was related to the acquisition of the Super Computer by Brazil called Santos Dumont and hosted at LNCC. It was built by a French company as a consequence of an agreement between both Brazilian and French government at the time of the french president Sarkozy and his Brazilian homologue Lula. The idea was that this Super Computer would encourage the research collaboration between the two countries and there will be technology transfer for the usage of the super computer platform. The HOSCAR project included Inria and different Brazilian groups (LNCC, UFRJ/COPPE, UFRGS,…).
Then, we had the SciDISC project. It was about scientific workflow and data intensive computing. We wanted to develop the research to understand how we could support what we call in-silico research.
This project had to do with the development of systems that would be implemented to scale to the cloud and do that efficiently. It was coordinate by Prof. Marta Mattoso from COPPE/UFRJ. It was followed up by the MUSIC project which was about multi-site cloud, which is the idea that scientific workflows could run distributedly on different clouds.
HPDaSc, the current project that you mentioned, which started this year, is also a project that I coordinate and it includes a group from UFRJ/COPPE with two professors (Prof. Marta Mattoso and Prof. Alvaro Coutinho), UFF in Niteroi (Prof. Daniel de Oliveira and Pr. Aline Paes), and CEFET-RJ (Prof. Eduardo Ogazawara and Prof. Rafaelli Coutinho).
This project is about running data analysis in High Performance Computer Systems (such as the Santos Dumont Super Computer). The idea is to use those machines to do data-driven research. As any state-of-the-art research project, it includes artificial intelligence (AI): we try to understand how AI could interpret this huge amount of data that is now available, using HPC systems. We are kind of empowering research by including AI components to this process.
What do you think this collaboration has brought to you? And to LNCC?
Well, this is the most relevant collaboration that we have, in many directions. First of all, it is a long lasting and continuing one that started twenty years ago. We have been obtaining new funding for each one of those projects. As soon as one finishes, we manage to have another one following up. That maintains the collaboration and the exchange of professors and students between Brazil and France. It enables us to somehow have an entrance in the European research community as well.
Inria and LNCC are very similar, in the way we are structured and in the way we work: we both are research centers that is not an university, focusing on computational models. This makes us kind of sibling institutions in both countries.
For LNCC, as a result of this long last collaboration, we have plenty bilateral collaborations: Artur Ziviani is a colleague which shares collaboration with Aline Vieira (Tribe/INRIA) and Eric Fleury, Frederic Valentin which has an International Chair position at Inria, Pr. Ana Tereza Vasconcelos who works with bioinformatics has also long last ties research with Inria.
One important mark of our collaboration was also to bring VLDB to Brazil in 2018. It’s a big conference in data management (the focus of my group, Pr. Patrick Valduriez and Pr. Esther Pacitti’s) that we managed together to bring to Brazil, which didn’t happen since 1975, I guess. I think it was a huge effort that was made possible through our cooperation.
What would be the next step of this collaboration, since your point of view?
The establishment of an International Lab is one of those. At this point we are signed a Memory of Understanding between LNCC and INRIA for the discussion around its structuring, and we have to transform that into a concrete virtual lab. It’s important to structure this initiative including the communication with the different funding agencies in Brazil, which are spread among different states and organizations. I think by having an International Lab we can structure the communication between Inria and funding agencies, but it would also enable us to focus on some research interests (such as AI).
Obviously, we have to continue having exchanges of students because it widely contributes to make the collaboration fruitful: during their master or PhD, the students implement the collaboration in terms of practical work.
I am very positive about the collaboration, that it will continue to be successful, long-lasting, and productive.