Academic seminar on AI and machine learning
Philip Thompson, professor at FGV EMAp, was the speaker on Wednesday (2)
3/4/2025
Philip Thompson, associate professor at FGV EMAp – School of Applied Mathematics, PhD and master’s degree from IMPA, was the guest speaker at the academic seminar on Wednesday (2). Organized by Professor Ueverton Souza, the event took place in IMPA Tech’s main classroom and brought together students from the classes of 2024 and 2025, reinforcing the collaboration between educational institutions in Rio de Janeiro.
“This exchange between researchers from different institutions is vital for strengthening science in our state and in the country. It’s important for students to have contact with the research carried out inside and outside IMPA, not least so that they can glimpse possibilities for future studies and different techniques that can be used for the same problem,” said Ueverton.
Under the title “Machine learning, the stochastic gradient method and corrupted data”, Thompson presented the students with research carried out in the field of machine learning.
Machine learning is an area of Artificial Intelligence that seeks to build algorithms capable of “learning” from data and then making predictions and decisions based on the information provided. This field is widely used in both Computer Science and Data Science.
“The area of machine learning is closely related to the areas of Statistics and Probability. In general, the “best” way of learning involves minimizing a cost function subject to constraints that suit the relevant characteristics of the desired model. One example is models described by neural networks. Thus, machine learning is closely related to the area of applied mathematics called Stochastic Optimization, which seeks to find function minima efficiently,” said Thompson.
Student Yuri da Silva, who came from Uberaba (MG) to start his studies at IMPA Tech, was excited about the meetings. “From the lecture, I had some ideas for applying machine learning, an area I didn’t know about. I think the seminars are great for opening up our horizons to new areas, even if we haven’t had much contact with the subject yet.”
Gabriel Vieira, from Limoeiro (PE), is a veteran of the course and was encouraged by Philip Thompson’s introduction. “I found the seminar very interesting. Philip seems to be very good at what he does. I was studying machine learning before and it’s a subject we’re going to see next term. It was great to be able to learn a bit more.”
Gabriel also highlighted the multidisciplinary nature of the seminars, which are opportunities to understand the applications of the studies. “We get to see a bit of each area applied to what we’re studying in a more comprehensive and practical way. We’ve already had seminars on biology applied to computing, music and mathematics, and today it was machine learning. It’s really cool!” concluded the student.
The academic seminars began in January of this year and have already been attended by several speakers in Maravelley Port. Among them were IMPA researchers, IMPA Tech professors and guests from other institutions, such as the Fluminense Federal University (UFF) and FGV EMAp – School of Applied Mathematics. This was the fifth meeting held in 2025.
