News

News

NOVA FCT researchers were awarded Best Paper at Marguerite Frank Prize

14-03-2024

Tiago Dias, a PhD student in Mathematics at NOVA FCT, and Paula Amaral, his supervisor, were awarded the Marguerite Frank Award for their paper "A classification method based on a cloud of spheres". Paula Amaral is a Professor in the Department of Mathematics at NOVA FCT and a researcher at NOVA MATH.

In this award-winning paper, a new supervised classification method based on a cloud of spheres is proposed. This process allows for the identification of non-linear and non-convex data structures. The classification process and method developed are suitable for anomaly detection and aim to offer an alternative to black-box methods. This article also critically examines some of the social implications of non-transparent Machine Learning methods, with the issue of transparency serving as one of the motivations for developing this new proposal.

"It was with immense satisfaction that we received the news of this prestigious award, which has particular significance given the inspiring role of Prof. Marguerite Frank as a mathematician and as a woman. This recognition serves as great encouragement and motivation for both of us to continue our work with renewed vigor, aiming to contribute to the advancement of science guided by the values that should guide humanity," said the researchers.

The EURO Journal on Computational Optimization is one of the publications of the Association of European Operational Research Societies, which publishes research articles and tutorials that explore the many areas in which Optimization and Computer Science are closely linked. The common element in all the contributions chosen for publication in the journal is innovation in areas where computational methods are used to solve complex optimization problems.

The name of the prize is a tribute to Marguerite Straus Frank, whose outstanding discoveries in the area of simple Lie algebra and the optimization of quadratic problems gave rise to the well-known Frank-Wolfe method.