Coloquio Primavera 2025
Tomás Teijeiro (BCAM)
AI – The mathematics of self‑deception
| Viernes, 25 de abril de 2025, 12:00h., Módulo 17, Aula 520 | |
| Cartel |
Abstract:The impact of "Artificial Intelligence" in science and technology in recent years has been enormous, giving rise to a bubble of overhyped claims and promises that heavily influence the research landscape in fields as diverse as chemistry, medicine or energy, to name just a few. However, it is increasingly evident that there is a dissonance between what AI does and what AI gurus tell us it can do.
In this talk, we will show why AI, and more specifically Machine Learning, is a perfect tool for self-deception, even if we are extremely careful with our experimental methodology. We will discuss the mathematical roots that can lead to overestimated performance results, and illustrate how they appear subtly in more real and "top-level" scientific studies than one would expect.
Coloquio Primavera 2024
Gérard Biau (Sorbonne Université)
Deep residual networks and differential equations
| Viernes, 8 de marzo de 2024, 12:00h., Módulo 17, Aula 520 | |
| Enlace online: youtube.com/live/WCPW9c5S1p8 | |
| Cartel |
Abstract: Deep learning has become a prominent approach for many applications, such as computer vision or neural language processing. However, the mathematical understanding of these methods is still incomplete. A recent approach is to consider neural networks as discretized versions of differential equations. I will first give an overview of this emerging field and then discuss new results on residual neural networks, which are state-of-the-art deep learning models.