How would you describe your work in 90 seconds?

I work at the intersection of Artificial Intelligence (AI) and Neuroscience. What we are trying to do in the field is to understand the principles of brain computation. There are many open mysteries on how the brain works: How does memory or reasoning emerge from electrical neuron activity? But we live in exciting times where brain imaging and computational theories of AI enable us to tackle this question in new ways. My research group analyzes brain activity recorded from cutting-edge technology, typically brain activity from a mouse brain. Still, it could be a human brain when the data is available. We then try to model these recordings with computational models to look for principles that explain what we see in the recordings, think of looking for the “equations of the brain”. When we believe that we finally understand a functional principle of brain computation, we then try to have a pragmatic and constructive approach to illustrate what it means for biology and computer science. Sometimes, it means that we will work with biologists to explain unknown brain phenomena, or with computer architects to build so-called neuromorphic systems that are more brain-like and innovatively emulate intelligence.

How did you get in touch with informatics?

This started with my dad, who was born in 1941. He was one of the very early Computer Scientists, and during his military service, he was sent from France to Brazil. There, he built the first computer in Salvador de Bahia, and at the time, the computer had no screen and weighed 300kg. Throughout his life, he kept learning about computers, and he continues today (he is 84, but he has a smartphone, of course). So, at home, we always had weird old and newer computers, and we were constantly playing around with them. If you grow up with computers all around you, they become familiar. As a teenager, I played video games, participated in online teams, and spent much time on my computer. Eventually, I also wanted to create my own website; I was probably 14. I thought it would be cool to have a chat platform or a website for the gaming team, and my friends from school. So, I learned PHP and HTML on the internet. Then a friend invited me to join a high school rock band, and before I could play anything with a guitar, I naturally found myself behind the computer to record the band and plan another website. So it started like that, and now it has become natural for me to do scientific programming for hours non-stop like I played games as a teenager.

Where do you see the connection between your work and everyday life?

AI has become part of everyday life; I have friends who talk to their favorite AI daily as a personal advisor, and I use it daily to code. I feel privileged that through teaching and research, I can keep up with these AI technologies and stay at the forefront of the technology. Understanding the brain is fundamental research and not practical for everyday life. But somehow, brain-like computing is thought to be a good model for sustainable AI. Soon, we will have many intelligent devices running around us when we eat and drink coffee. Still, the energy consumption of AI algorithms might become a problem for our batteries and electricity consumption. The vision of neuromorphic computing is that small chips, with tiny digital neural network structures inspired by the brain, could power applications on your mobile phone or run day-to-day tasks for you. Many researchers believe this alternative brain-like chip design could make AI more power-efficient. The field is not quite there yet, but I like to build stuff, so I took steps in that direction. For example, we created a mobile application for musical chord recognition, called Chord AI. You give the audio, and it will transcribe the chords for you and help you play-along for guitar, piano, or ukulele. This seemed helpful as I like to play music as a hobby. There was a real technological achievement there, though, because the AI algorithm runs directly on the chip in your phone. This is what you can do with today’s technology, and I believe it will become more common to use AI for all kinds of things, and the models will become even more capable. I want to continue being an actor in this development.

What makes you happy in your work?

I love to demystify things, making complex things simple. I’m curious about how things work and enjoy finding logical explanations. I’m looking for truth and for understanding the world. The brain is one of the more mysterious things in science at the moment. It may appear mystical and infinitely complex at first, but when scientists dig carefully, they can find a complete, coherent story and sometimes a simple explanation. At that point, I feel this “aha” moment and want to share these discoveries with my peers, students, and friends. This is also what drives me when I’m teaching, but also when I’m writing research papers. Writing papers has become one of the most exciting parts of my work. It’s a creative process where you plug out from the world for a few days; the beginning can be hard, but at some point, much like it is with programming, you get into the zone, and then you don’t want to stop. This is when your thoughts crystallize, and you write paragraph after paragraph (with lots of typos, but it doesn’t matter; you will correct them eventually): this is a very delightful moment. To me, academic writing is like a logical puzzle that reveals whether you hide the truth or are dishonest with your reasoning. And when it works, it’s about the joy of sharing the simple inner mechanism of something that is otherwise complex and mysterious. It wasn’t always like that, though; it’s only towards the end of my PhD that I discovered joy in the writing process.

Why do you think there are still so few women in computer science?

It’s a problem; we need all the computer scientists and programmers we can, and we need more women in computer science. I see no reason why women wouldn’t be able to do that job better than men. Today, only 15 to 30% of students in computer science are women; the proportion may be a bit higher in some applied fields. The reason behind this is a mystery to me. Since we see this imbalance early at university, it seems like young women already prefer to study other things. We could challenge the stereotypes before that. Teenage girls could be more encouraged to have fun building things with a computer. It’s both about showing that being a nerd is not just fun for boys, and also showing that other skills are involved, like artistic creativity, and social understanding. This is just one aspect, and it’s a complex problem without a simple solution, but having this debate is useful for sure. Later in the scientific career, I believe that some policies can encourage diversity without compromising excellence and scientific progress.

Guillaume Bellec is Assistant Professor at the Research Unit Machine Learning at TU Wien Informatics. His current project, AI and neuroscience, explores how Artificial Intelligence can be leveraged to decode the functioning of the brain using electrophysiological recordings. The project is funded by a WWTF VRC grant and is set to run until 2033.

Discover the whole #5QW series.

Curious about our other news? Subscribe to our news feed, calendar, or newsletter, or follow us on social media.