Quentin Berger
Professeur, Université Sorbonne Paris Nord (aka Paris 13)
Laboratoire d'Analyse, Géométrie et Applications (LAGA),
Équipe Probabilités et Statistique.
Membre junior de l'Institut Universitaire de France (IUF).
Membre élu du Comité National (Section 01).
Adresse/Address:
LAGA, Université Sorbonne Paris Nord,
99 Av. J-B Clément, 93430 Villetaneuse, France
Mon bureau/my office: D301
email: quentin.berger(at)univ-paris13.fr
Mes profils arXiv, HAL, Orcid et Google Scholar.
Domaine de recherche: Probabilités et Mécanique Statistique, EDP Stochastiques singulières.
Mots-clés: systèmes désordonnés et influence du désordre, modèles de polymères aléatoires, transition de phase et phénomènes critiques, limites d'échelles et EDPs stochastiques, marches aléatoires, processus de renouvellements, graphes aléatoires, percolation de premier et de dernier passage, etc.
Some personal comments on our work on the critical 2D KPZ process. Recently, I got involved in some (experimental) collaborative work, which led to the rigorous construction of a nontrivial solution to the ill-defined 2D KPZ equation at criticality. The results build upon a decade's collective effort and several recent breakthroughs, but were obtained through extensive interactions with generative AI. It still took us a lot of time and work to polish off the manuscript, as we wanted to offer something valuable to the community.
Quentin Berger, Francesco Caravenna, Jeremy Clark, Alexander Dunlap, Yu Gu, Ziyang Liu, Shuta Nakajima, Rongfeng Sun, Li-Cheng Tsai, Zuodi Xie, Nikos Zygouras,
The critical 2D KPZ process,
arXiv:2610.31614 [math.PR]
Let me make some more personal comments on why I see this as a positive message, in response to the current AI frenzy:
  • It questions the notion of authorship of an article. I believe that we cannot individually claim credit for the result in itself: it instead highlights the collective effort of our community over several years.
  • It stresses the importance of collaborative work. The multi-authorship shows that it is still possible (and desirable) to set aside the competition in order to exchange ideas and improve our common understanding.
The current pace at which results are released (not to mention the batch releases) is clearly not sustainable and puts a great toll on mathematicians mental health, especially PhD students and postdocs. Many of us are still trying to figure out how to restore some serenity in our work, how to train younger mathematicians and offer them a desirable career. Questioning the idea of credit in the age of AI, putting more value on the understanding and reaffirming the importance of human collaborative work (either with or without AI assistance) may be some first steps.

Notes de cours/Lecture notes:

Livre d'Introduction aux probabilités

Avec Francesco Caravenna et Paolo dai Pra. Quelques exercices et exemples supplémentaires qui nous paraissent intéressants, toujours dans l'esprit du livre.

Autres/Others:

Des articles grand public : sur le mélange de cartes, sur le paradoxe de Simpson.
Participation au projet ANR Local.

Événements/Events:

Événements à venir/Upcoming events:
Quelques événements passés/Some past events :
Sep. 2026: Random Interacting Systems, StatMech Summer School, Padova, Italie.
May 2026: Statistical Mechanics and Singular SPDEs, Singapour.
Jan. 2026: Random walks: applications and interactions, at CIRM.
Nov. 2025: Probability, Analysis and Mathematical Physics, Nanterre.
Nov. 2025: Rencontres de probabilités de Rouen.