Research by FGV students on AI presented at ICML 2024

Two students at Fundação Getulio Vargas’ School of Applied Mathematics (FGV EMAp) in Rio de Janeiro went to Vienna to attend the 41st edition of the International Conference on Machine Learning (ICML). The event, held last week, is recognized as one of the world’s leading academic conferences dedicated to artificial intelligence and machine learning. As well as lectures and workshops, the conference includes a session for exhibiting and presenting scientific papers.

Daniel Csillag and Tiago da Silva Henrique submitted papers to the rigorous double-blind review process carried out by the event’s organizers and they were among researchers whose work reflects the most innovative and crucial advances in current artificial intelligence and machine learning research.

“This conference is internationally recognized, and having our work published there is an important achievement that demonstrates that we are competing globally with renowned researchers,” said Tiago.

Daniel wrote his paper in conjunction with Claudio Struchiner and Guilherme Tegoni Goedert, professors at the school. The article, titled “Generalization Bounds for Causal Regression: Insights, Guarantees and Sensitivity Analysis,” provides theoretical guarantees for causal regressions.

“We set out to evaluate and establish which of the existing algorithms can reliably and rigorously retrieve causal information under simple and applicable assumptions. In this way, we ensure that the predictions made by these algorithms reflect real causal relationships, rather than mere correlations,” explained Daniel.

Tiago produced his paper, “Embarrassingly Parallel GFlowNets,” in partnership with Luiz Max Carvalho and Diego Mesquita, researchers at FGV EMAp, as well as Amauri Souza and Samuel Kaski, researchers at Aalto University in Finland. In the article, they developed an algorithm to perform parallel Bayesian inference on distributions over a finite number of variables.

Young researchers in the spotlight

Tiago’s academic career has been marked by excellence since the days when he lived in Florianópolis and attended a public sector elementary school. After doing extremely well in the Brazilian Public School Mathematics Olympiad (OBMEP), he decided to pursue higher education at FGV EMAp, where he found a pioneering course that combined computing, mathematics and data analysis. “I was really impressed not only with the program’s innovation, but also the opportunity to have classes with a distinguished team of professors,” he says.

Now doing a master’s program, Tiago is also enrolled in some doctoral courses, which he believes has provided him with a solid basis for carrying out high-quality research. “One of the most valuable aspects of FGV EMAp is having this opportunity to access graduate courses while doing undergraduate studies,” he says.

Daniel, who is from Sao Paulo, points out that the school actively promotes contact between its students and cutting-edge research from all over the world. “FGV EMAp offers a great environment for high-quality academic collaboration and production. Events like this provide an excellent opportunity to publish and disseminate our work,” he adds.