Publications

@inproceedings{verine2025changetheloss,
  title = {Improving Diversity in Language Models: When Temperature Fails, Change the Loss},
  author = {Vérine, Alexandre and Bronnec, Florian Le and Zheng, Kunhao and Allauzen, Alexandre and Chevaleyre, Yann and Negrevergne, Benjamin},
  booktitle = {ICML},
  year = {2025},
}

@inproceedings{tanner25a,
  title = {A High Dimensional Statistical Model for Adversarial Training: Geometry and Trade-Offs},
  author = {Tanner, Kasimir and Vilucchio, Matteo and Loureiro, Bruno and Krzakala, Florent},
  booktitle = {AISTATS},
  year = {2025},
}

@inproceedings{dandi25rmt,
  title = {A Random Matrix Theory Perspective on the Spectrum of Learned Features and Asymptotic Generalization Capabilities},
  author = {Dandi, Yatin and Pesce, Luca and Cui, Hugo and Krzakala, Florent and Lu, Yue and Loureiro, Bruno},
  booktitle = {AISTATS},
  year = {2025},
}

@inproceedings{troiani25fundamental,
  title = {Fundamental computational limits of weak learnability in high-dimensional multi-index models},
  author = {Troiani, Emanuele and Dandi, Yatin and Defilippis, Leonardo and Zdeborova, Lenka and Loureiro, Bruno and Krzakala, Florent},
  booktitle = {AISTATS},
  year = {2025},
}

@inproceedings{verine2025improvingdiscriminatorguidancediffusion,
  title = {Improving Discriminator Guidance in Diffusion Models},
  author = {Vérine, Alexandre and Inane, Ahmed Mehdi and Bronnec, Florian Le and Negrevergne, Benjamin and Chevaleyre, Yann},
  year = {2025},
  booktitle = {ECML},
}

@article{soletskyi2025theoretical,
  title = {A theoretical perspective on mode collapse in variational inference},
  author = {Soletskyi, Roman and Gabri{\'e}, Marylou and Loureiro, Bruno},
  year = {2025},
  journal = {MLST},
  url = {https://iopscience.iop.org/article/10.1088/2632-2153/adde2a},
}

@article{defilippis2025optimal,
  title = {Optimal Spectral Transitions in High-Dimensional Multi-Index Models},
  author = {Defilippis, Leonardo and Dandi, Yatin and Mergny, Pierre and Krzakala, Florent and Loureiro, Bruno},
  year = {2025},
  journal = {arXiv},
}

@article{fan2025dynamical,
  title = {Dynamical mean-field analysis of adaptive Langevin diffusions: Propagation-of-chaos and convergence of the linear response},
  author = {Fan, Zhou and Ko, Justin and Loureiro, Bruno and Lu, Yue M and Shen, Yandi},
  year = {2025},
  journal = {arXiv},
}

@article{fan2025dynamicalb,
  title = {Dynamical mean-field analysis of adaptive Langevin diffusions: Propagation-of-chaos and convergence of the linear response},
  author = {Fan, Zhou and Ko, Justin and Loureiro, Bruno and Lu, Yue M and Shen, Yandi},
  year = {2025},
  journal = {arXiv},
}

@article{vilucchio2025existence,
  title = {On the existence of consistent adversarial attacks in high-dimensional linear classification},
  author = {Vilucchio, Matteo and Zdeborov{\'a}, Lenka and Loureiro, Bruno},
  year = {2025},
  journal = {arXiv},
}

@article{arnaboldi2025asymptotics,
  title = {Asymptotics of SGD in Sequence-Single Index Models and Single-Layer Attention Networks},
  author = {Arnaboldi, Luca and Loureiro, Bruno and Stephan, Ludovic and Krzakala, Florent and Zdeborova, Lenka},
  year = {2025},
  journal = {arXiv},
}

@inproceedings{2024_cyffers_performative,
  title = {Optimal Classification under Performative Distribution Shift},
  author = {Cyffers, Edwige and Pydi, Muni Srinivas and Atif, Jamal and Cappé, Olivier},
  booktitle = {NeurIPS},
  year = {2024},
}

@inproceedings{Defilippis2024,
  title = {Dimension-free deterministic equivalents and scaling laws for random feature regression},
  author = {Defilippis, Leonardo and Loureiro, Bruno and Misiakiewicz, Theodor},
  booktitle = {NeurIPS},
  year = {2024},
}

@inproceedings{cui24d,
  title = {Asymptotics of feature learning in two-layer networks after one gradient-step},
  author = {Cui, Hugo and Pesce, Luca and Dandi, Yatin and Krzakala, Florent and Lu, Yue and Zdeborova, Lenka and Loureiro, Bruno},
  booktitle = {ICML},
  year = {2024},
}

@inproceedings{arnaboldi24a,
  title = {Online Learning and Information Exponents: The Importance of Batch size \& {T}ime/{C}omplexity Tradeoffs},
  author = {Arnaboldi, Luca and Dandi, Yatin and Krzakala, Florent and Loureiro, Bruno and Pesce, Luca and Stephan, Ludovic},
  booktitle = {ICML},
  year = {2024},
}

@inproceedings{verine2024exploringprecisionrecall,
  title = {Exploring Precision and Recall to assess the quality and diversity of LLMs},
  author = {Vérine, Alexandre and Bronnec, Florian Le and Negrevergne, Benjamin and Chevaleyre, Yann and Allauzen, Alexandre},
  year = {2024},
  booktitle = {ACL},
}

@inproceedings{schroder2024deterministic,
  title = {Asymptotics of Learning with Deep Structured (Random) Features},
  author = {Schröder, Dominik and Dmitriev, Daniil and Cui, Hugo and Loureiro, Bruno},
  booktitle = {ICML},
  year = {2024},
}

@inproceedings{clarte2024bootstrap,
  title = {Analysis of Bootstrap and Subsampling in High-dimensional Regularized Regression},
  author = {Clart\'e, Lucas and Vandenbroucque, Adrien and Dalle, Guillaume and Loureiro, Bruno and Krzakala, Florent and Zdeborov\'a, Lenka},
  booktitle = {UAI},
  year = {2024},
}

@article{gerace2024gaussian,
  title = {Gaussian universality of perceptrons with random labels},
  author = {Gerace, Federica and Krzakala, Florent and Loureiro, Bruno and Stephan, Ludovic and Zdeborov{\'a}, Lenka},
  year = {2024},
  journal = {PRE},
}

@article{dandi2024two,
  title = {How two-layer neural networks learn, one (giant) step at a time},
  author = {Dandi, Yatin and Krzakala, Florent and Loureiro, Bruno and Pesce, Luca and Stephan, Ludovic},
  year = {2024},
  journal = {JMLR},
}

@article{adomaityte2024high,
  title = {High-dimensional robust regression under heavy-tailed data: Asymptotics and universality},
  author = {Adomaityte, Urte and Defilippis, Leonardo and Loureiro, Bruno and Sicuro, Gabriele},
  year = {2024},
  journal = {JSTAT},
  url = {https://iopscience.iop.org/article/10.1088/1742-5468/ad65e6},
}

@article{arnaboldi2024escaping,
  title = {Escaping mediocrity: how two-layer networks learn hard generalized linear models with SGD},
  author = {Arnaboldi, Luca and Krzakala, Florent and Loureiro, Bruno and Stephan, Ludovic},
  year = {2024},
  journal = {arXiv},
}

@article{vilucchio2024geometry,
  title = {On the Geometry of Regularization in Adversarial Training: High-Dimensional Asymptotics and Generalization Bounds},
  author = {Vilucchio, Matteo and Tsilivis, Nikolaos and Loureiro, Bruno and Kempe, Julia},
  year = {2024},
  journal = {arXiv},
}