Using Academic Expertise in Applied Mathematics for Student Data Mining and Machine Learning Competitions (Thibaut Lunet, Hamburg University of Technology, Germany)

16.11.2021 14:00 – 15:00

Since 2019, the University of Geneva participates in a German Student competition, the Data-Mining Cup, which takes place each year and challenges university students from all around the world on a specified task inspired from real life applications, requiring knowledge in data mining, statistics and modeling.
In 2019, a combined team with ETH Zurich led by Renaud Rivier (PhD student) proposed a solution for a fraud detection problem, and got to the second place.
In 2020, reinforced by two postdoctoral fellows (Raphael Ducatez and Thibaut Lunet), the team achieved the third place for a website sales prediction task.
Finally in 2021, two teams of students from the mathematics department, lead by Renaud Rivier and Thibaut Lunet,
worked on a suggestion algorithm for online book selling, and one of them got to the first place.

In this seminar, we will briefly present the scientific aspect of those competitive tasks, and the numerical methods developed by the UniGe teams to solve them.
The presentation will also cover what students and members of teaching staff can gain (beside prize money) when participating in such competitions : the opportunity to teach and learn new aspects of applied mathematics,
the experience of team work focused on one common objective, the use of academical knowledge and experience for real life applications ...
We will also give a quick overview of current competitions to which university students can participate, eventually with the help of professors in mathematics and computer science.
Finally, a parallel will be made with the CAMAS, a consulting platform of the Mathematical Department intending to provide applied mathematical support and solutions for other departments of the University of Geneva.

Lieu

Conseil Général 7-9, Room 1-05, Séminaire d'analyse numérique

entrée libre