BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Summer school "Deep Learning and applications"
DTSTART:20250825T041000Z
DTEND:20250829T140000Z
DTSTAMP:20260423T160700Z
UID:indico-event-14421@indico.math.cnrs.fr
CONTACT:summerschool2025@math.unistra.fr
DESCRIPTION:This summer school offers an introduction to deep learning and
  its various applications\, such as physical modeling and the formalizatio
 n of mathematics. It covers several recent techniques\, such as transforme
 rs or Large Language Models\, as well as mathematical methods to analyze t
 heir learning process.\nThe summer school is composed of five lectures and
  three talks. Each lecture (2x1h30) will be complemented by 2 hours of tut
 orials. It is primarily intended for the students of the graduate program
  "Mathematics and interactions: research and interactions" of the Univers
 ity of Strasbourg but is also accessible to any interested PhD student or 
 researcher. As space is limited\, priority will be given to PhD students. 
 This event is supported by the Interdisciplinary Thematic Institute IRMIA
 ++. \nIt will take place from 25 to 29 August in the IRMA conference room
  at the University of Strasbourg.   \n\nLectures \nIntroduction to deep
  learning\, Antoine Deleforge (INRIA\, Université de Strasbourg) \nFor
 malizing mathematics with Large Language Models\, Marc Lelarge (INRIA\, 
 ENS) \nTransformers and flows in the space of probability measures\, Dom
 enech Ruiz I Balet (Université Paris Dauphine) \nIn this series of lect
 ures\, we will investigate mathematical properties of transformers. We wil
 l see them as partial differential equations and we will discuss mathemati
 cal properties of them\, such as universal approximation\, clustering and 
 other phenomena. The lectures are complemented with a coding session.\nDee
 p Reinforcement Learning and AI for Symbolic Mathematics\, Wassim Tenachi
  (MILA\, Université de Montréal) \nIn this course\, I will introduce d
 eep reinforcement learning (RL) - the machine learning approach behind man
 y of the impressive videos we have all seen of agents mastering video game
 s - while highlighting its much broader potential. The session is designed
  to be intuitive and accessible\, aiming to give students a solid grasp of
  what RL is capable of\, and how it works. I will cover the fundamental pr
 inciples and real-world use cases of RL\, a paradigm where agents learn th
 rough trial and error in simulated or real environments\, without necessar
 ily having access to explicit gradients. Instead\, they learn by approxima
 ting the gradients required to train neural networks. To ground these idea
 s\, I will present a case study from my own research in AI for symbolic ma
 thematics\, where machine learning models assist in tasks like theorem pro
 ving\, symbolic equation discovery\, and even replicating aspects of empir
 ical sciences - such as physics or astrophysics - by uncovering analytical
  expressions that model observed data: a field known as symbolic regressio
 n. We will in particular explore how numerical models like neural networks
  can be interfaced with symbolic mathematical structures\, and how such hy
 brid systems can learn to achieve specific reasoning goals. In the practic
 al session\, we will dive into some simple yet fun toy examples to better 
 understand both reinforcement learning and symbolic regression. These hand
 s-on activities are meant to give students a foundation they can build on 
 in future projects or research.\nTime series reasoning with language model
 s\, Svitlana Vyetrenko (Université de Strasbourg) \nTalk \nMoving fro
 m optimal control to continuous-time reinforcement learning using tools fr
 om SciML\, Alena Shilova (INRIA Saclay)\n\nhttps://indico.math.cnrs.fr/eve
 nt/14421/
IMAGE;VALUE=URI:https://indico.math.cnrs.fr/event/14421/logo-2529263734.pn
 g
LOCATION:Salle de Conférences (IRMA)
URL:https://indico.math.cnrs.fr/event/14421/
END:VEVENT
END:VCALENDAR