Artificial intelligence for research and training

Article en français

In a society where artificial intelligence (AI) is becoming more and more prevalent every day, understanding its challenges and mastering its applications is becoming essential.
Alumnus of Sorbonne University and research engineer at SCAI (Sorbonne Center of Artificial Intelligence), Baptiste Gregorutti is an expert in data sciences. Between training, research projects and awareness of AI, he gives us his insights on the importance of training the university community on these issues.

Can you tell us about your professional background?

Baptiste Gregorutti: I obtained a master's degree in mathematics at the Faculty of Sciences and Engineering of Sorbonne University, then completed a Cifre(1) thesis at the probability, statistics and modeling laboratory, under the direction of Gérard Biau, now director of SCAI, by Philippe Saint Pierre and Bertrand Michel. My thesis, defended in 2015, focused on the analysis of black boxes for aviation safety at Safety Line. I continued to work for them until 2020, focusing in particular on the economics of fuel consumption in aeronautics. I grew with the company, going from a small start-up of 2 people to a team of 30 employees. During this period, I developed skills in technical project and team management.

I then joined a consulting company founded by a data science researcher as technical leader and then scientific director. This led me to work on topics related to cybersecurity. After a short professional break, I set up my own business and reconnected with my former thesis director, Gérard Biau.

What is your current mission within SCAI?

B. G.: As a research engineer at SCAI, I am responsible for several missions. Part of my work concerns training, where I organize workshops for different audiences: doctors, students, high school students. I familiarize them with artificial intelligence and computer science by offering them practical workshops to put them in the shoes of a data scientist.

Furthermore, I work on research projects in collaboration with scientists from the university who have data, but do not know how to use it. For example, I am currently supervising a biology intern who is looking to create a machine learning model to identify bumblebee species from wing shape detected in photos. I am also involved in a project at Saint-Antoine hospital to understand why some women have repeated miscarriages. We also have a project in preparation with musicologists. My role is to guide these technical projects and help build effective tools.

Who are the courses in AI and data sciences that you are setting up for?

B. G.: The courses that I give in AI and data sciences are open to a wide audience. I work with biologists, chemists, literary scientists, doctors, across the three faculties of Sorbonne University.
In addition to continuing education, I am involved in initial teaching by giving courses at the Institute of Statistics of Sorbonne University and at the Science Apprentice Training Center of Sorbonne University. I also support undergraduate math students on science and data subjects.

In your opinion, why is it essential today to have knowledge of AI and data sciences even when you do not work directly in this field?

B. G.: AI is omnipresent in our society. It goes well beyond conversational assistants like Siri which are often the representation we have of AI in the collective imagination. For example, it is used in image processing or photo enhancement directly in smartphones. It also plays a role in the analysis of texts in literature and the structuring of information.

Some people think AI is “magic,” but in reality it is a complex technology that has limitations and makes mistakes. It is also important to take into account its environmental impact because certain applications, such as ChatGPT, are very energy intensive. It is therefore essential to demystify AI and understand its challenges to better understand these technologies in our daily lives and to be more critical of their use.

What are the next projects that you plan to carry out within Sorbonne University?

B. G.: We have several projects in sight. In the field of human and social sciences, one of the current projects concerns the development of a generative AI tool aimed at interacting with all the publications of researchers at Sorbonne University. The objective is to improve the management and exploitation of the vast quantities of available data, while making a new knowledge base available to the French-speaking and international community, and promoting our scientific advances.

At the same time, depending on requests and available resources, we will continue to support all members of the Sorbonne University community in their projects requiring AI and data science.
We are also working on improving the SCAI Education online course platform to understand artificial intelligence and learn its different uses and tools.

1: Industrial research training agreements