A series of weekly meetings on causality, led by mathematician and teacher Mag. Gabriel Illanes of IDATHA will be held until November 24.

Every Thursday morning, the seminar Causal Inference in Statistics, promoted by IDATHA, a member of the Pyxis ecosystem, together with the Mathematics Center of the Sciences Faculty of UDELAR, will be held. It is possible to participate in person or remotely.
Virtuality is one of the most chosen and with great advocacy, even universities from other parts of the world are already taking part.

It´s oversee it by the mathematician Gabriel Illanes, a fundamental part of the IDATHA team and a professor in the Engineering Faculty. With Gabriel and his knowledge as a common link, interesting conversations with ideas arose between IDATHA’s CEO Sebastián García and mathematician Ernesto Mordecki, professor at UDELAR and a reference in the area of probability and statistics, who also was Illanes’ master’s and doctoral tutor.

Ernesto was a member of the Honorary Scientific Advisory Group (GACH) during the covid-19 health emergency, so he had already used and studied causality tools to answer some of the questions that arose during the pandemic.

«And so, we discussed the idea between IDATHA and Mordecki as a referent of the Mathematics Center, with the motivation to study together causality and the applications of this novel methodology in a hybrid modality, so that everyone can participate», said Gabriel Illanes, whom we interviewed to know all the details.

For the content of this seminar, we took one of the most recognized books on the subject: Causality, by philosopher and computer scientist Judea Pearl. Class by class, the different chapters are progressed through, following the book in order, which makes it easy for students to join at any time they want.

In general, there are great advantages of virtuality, but in this case, they are multiplied, because those interested can start at any time, download the book and access the recorded meetings without missing anything.

«We have a fantastic book that we follow, so it’s quite straightforward; we choose what subjects are coming up session by session, and whomever is interested may read it ahead and then remark on it together, perform exercises, and discuss the intuition of the problem.»

There aren’t many prerequisites for studying the topic. It is important to have a basic understanding of probability or to have taken a probability course from the faculty.

According to Illanes, this format was designed «to be useful for faculty students to obtain credits and study it, but also to be friendly and reasonable for people from other places, other faculties, other industries, IDATHA members or clients, to join and follow the seminar. Part of the interesting thing about this is that it is a heterogeneous seminar from all aspects».

Gabriel Illanes is a young Uruguayan mathematician who has been teaching at the faculty for 15 years. He has a master’s degree in Mathematical Engineering and is about to receive his Ph.D. in Mathematics from the University of the Republic. His talent presented him with new challenges and the opportunity to study in European countries such as France and Sweden, where he acquired enriching knowledge that he now applies in our country.

IDATHA met Gabriel and from there he was motivated to join the company as part of the team. Since the beginning of this year, Gabriel, together with the company’s decision-makers, have been working directly in the area on several projects.

«We are already working and presenting solutions to some companies and projects of current clients to evaluate if these tools are applicable,» concluded the mathematician.

Causality

Causality is an area that started to become relevant about 15 years ago but is gaining momentum worldwide as it is of great help to machine learning in general.

In statistics or probability, it is often said that the fact that the different variables being measured are correlated does not mean that one causes the other. Correlation does not imply causation. For real-life problems, it is important to be able to conclude things about causality.

What happens if we condition one variable, what happens to the independence of the others? This study allows us to conjecture things about what’s going on and test them.

«We are seeing that there are some structures that are equivalent to others in the sense that if we don’t rule out one, we can’t rule out another. So we conclude that it is not a perfect methodology, but it does help us to improve a little bit the intuition that we could come up with,» Illanes added.

«There are no experts in this subject, we are all learning. The idea is to understand some causal structures. Here we are trying to help students to conclude correct things from the data.»

Today at IDATHA there is a working group that is studying and thinking about the application of causal structures in projects in our country.

If you want to know more about causality, we share with you the article published by Gabriel Illanes, which is available on IDATHA’s blog.