News

MILES was represented at IJCAI 2026 with the paper

• The Communication Complexity of Instant-Runoff Voting by Élie de Panafieu, François Durand and Jérôme Lang

The following papers co-authored by MILES members have been accepted to the 2026 International Conference on Machine Learning (ICML):

• Trading Complexity for Expressivity Through Structured Generalized Linear Token Mixing, by Erwan Fagnou, Paul Caillon, Blaise Delattre (MILES alumni) and Alexandre Allauzen,
• Certified Robustness under Heterogeneous Perturbations via Hybrid Randomized Smoothing, by Blaise Delattre (MILES alumni), Hengyu Wu, Paul Caillon, Wei Yang Bryan Lim and Yang Cao
• Optimal Regularization for Performative Learning by Edwige Cyffers, Alireza Mirrokni and Marco Mondelli

Congrats to all!

MILES will be represented at ICLR 2026 with the following papers:

• Scaling Direct Feedback Learning with Jacobian Alignment Guarantees, by Paul Caillon, Erwan Fagnou, Blaise Delattre (MILES alumni) and Alexandre Allauzen.
• Unified Privacy Guarantees for Decentralized Learning via Matrix Factorization, by Aurélien Bellet, Edwige Cyffers, Davide Frey, Romaric Gaudel, Dimitri Lerévérend and François Taïani.

Make sure to catch the authors in Rio!

MILES was represented at AAAI 2026 with the paper

• Nested Depth Search by Junkang Li, Tristan Cazenave, Swann Legras, Arthur Quéffelec and Véronique Ventos.

Edwige Cyffers is joining MILES as a CNRS researcher!

Edwige’s research interests include Privacy in Machine Learning (ML), Federated and Decentralized ML, as well asa Trustworthy ML. She is also very active in the Machine Learning and math communities, with a current involvement in the organization of EGMO 2026. Feel free to reach out to Edwige for more!

MILES was represented at ECAI 2025 in Bologna with the following paper:

• Forward Only Learning for Orthogonal Neural Networks of any Depth, by Paul Caillon, Alex Colagrande, Erwan Fagnou, Blaise Delattre (MILES alumni) and Alexandre Allauzen

MILES was represented at ECMLPKDD 2025 with the following papers:

• Gathering and Exploiting Higher-Order Information when Training Large Structured Models by Pierre Wolinski
• Improving Discriminator Guidance in Diffusion Model by Alexandre Vérine (MILES alumni), Ahmed Mehdi Inane, Florian Le Bronnec, Benjamin Negrevergne and Yann Chevaleyre
• Lattice Climber: Adversarial attacks for randomized mixtures of classifiers by Lucas Gnecco Heredia (MILES alumni), Benjamin Negrevergne and Yann Chevaleyre.

All had a great time in Porto!

Several papers featuring MILES team members have been accepted to the Forty-Second International Conference on Machine Learning (ICML 2025):

•  Wrapped Gaussian on the manifold of Symmetric Positive Definite Matrices by Thibault de Surrel, Fabien Lotte, Sylvain Chevallier and Florian Yger (MILES alumni);
• Improving Diversity in Language Models: When Temperature Fails, Change the Loss  by Alexandre Vérine (MILES Alumni), Florian Le Bronnec, Kunhao Zheng, Alexandre Allauzen, Yann Chevaleyre, and Benjamin Negrevergne.
• Exploring Large Action Sets with Hyperspherical Embeddings using von Mises-Fisher Sampling by Walid Bendada, Guillaume Salha-Galvan, Romain Hennequin, Théo Bontempelli, Thomas Bouabca and Tristan Cazenave.
• Optimizing Language Models for Inference Time Objectives using Reinforcement Learning by Yunhao Tang, Kunhao Zheng, Gabriel Synnaeve and Rémi Munos
• RLEF: Grounding Code LLMs in Execution Feedback with Reinforcement Learning by Jonas Gehring, Kunhao Zheng, Jade Copet, Vegard Mella, Taco Cohen and Gabriel Synnaeve
• PILAF: Optimal Human Preference Sampling for Reward Modeling by Yuzhen Feng, Ariel Kwiatkowski, Kunhao Zheng, Julia Kempe and Yaqi Duan.

Three papers accepted at AISTATS 2025 that will be held in May 2025 in Mai Khao, Thailand:

• Bridging the Theoretical Gap in Randomized Smoothing, by Blaise Delattre, Paul Caillon, Quentin Barthélémy (Foxstream), Erwan Fagnou and Alexandre Allauzen.
• Unveiling the Role of Randomization in Multiclass Adversarial Classification: Insights from Graph Theory, by Lucas Gnecco Heredia, Matteo Sammut, Muni Sreenivas Pydi, Rafael Pinot (Sorbonne Université), Benjamin Negrevergne and Yann Chevaleyre.
• Memorization in Attention-only Transformers, by Léo Dana (ENS Lyon and former MILES intern), Muni Sreenivas Pydi and Yann Chevaleyre.

Congrats to all!

Three papers featuring MILES team members accepted at ICLR 2025, to take place in Singapore at the end of April:

• Accelerated Training through Iterative Gradient Propagation Along the Residual Path, by Erwan Fagnou, Paul Caillon, Blaise Delattre, Paul Caillon and Alexandre Allauzen.
• SCOPE: A self-supervised framework for improving faithfulness in conditional text generation by Song Duong (Sorbonne Université & Criteo AI Lab), Florian Le Bronnec, Alexandre Allauzen, Vincent Guigue (AgroParisTech), Alberto Lumbreras (Criteo AI Lab), Laure Soulier (Sorbonne Université) and Patrick Gallinari (Sorbonne Université & Criteo AI Lab).
• What Makes Large Language Models Reason in (Multi-Turn) Code Generation?, by Kunhao Zheng, Juliette Decugis (Meta AI), Jonas Gehring (Meta AI), Taco Cohen (Meta AI), Benjamin Negrevergne and Gabriel Synnaeve (Meta AI).

Check them out at the conference!