2025 & 2026 selected papers from AI Lab (ICML, ICLR, AISTATS…)

2026

ICLR 2026 (4)

• Fused-Planes: Why Train a Thousand Tri-Planes When You Can Share? Karim Kassab, Antoine Schnepf, Jean-Yves Franceschi, Laurent Caraffa (LASTIG / IGN / Université Gustave Eiffel), Flavian Vasile, Jérémie Mary, Andrew I. Comport (CNRS / I3S / Université Côte d’Azur), Valérie Gouet-Brunet (LASTIG / IGN / Université Gustave Eiffel)
• Contextual Multi-Armed Bandits with Minimum Aggregated Revenue Constraints. Ahmed Ben Yahmed (CREST / Criteo AI Lab), Marc Abeille, Vianney Perchet (CREST / Fairplay / Criteo AI Lab), Hafedh El Ferchichi (CREST / Fairplay)
• On the Impact of the Utility in Semivalue-based Data Valuation. Mélissa Tamine (CREST / Fairplay / Criteo AI Lab), Benjamin Heymann, Maxime Vono, Patrick Loiseau (INRIA / Fairplay)
• Learning in Prophet Inequalities with Noisy Observations. Jung-hun Kim (CREST / Fairplay), Vianney Perchet (CREST / Fairplay / Criteo AI Lab)

AISTATS 2026 (4)

• On the Hardness of Reinforcement Learning with Transition Look-Ahead. Corentin Pla (CREST / Fairplay / Criteo AI Lab), Hugo Richard, Marc Abeille, Vianney Perchet (CREST / Fairplay / Criteo AI Lab)
• Sequential Off-Policy Learning with Logarithmic Smoothing. Maxime Haddouche (ENS Paris / Inria Sierra), Otmane Sakhi
• Unmixing Mean Embeddings for Domain Adaptation with Target Label Proportion. Alain Rakotomamonjy, Maxime Bérar (LITIS – Université de Rouen Normandie), Mokthar Z. Alaya (LMAC – UTC)
• Contextual Ranking and Matching: Optimal Regret under LST. Hafedh El Ferchichi (CREST / Fairplay), Matthieu Lerasle (CREST / Fairplay), Vianney Perchet (CREST / Fairplay / Criteo AI Lab)

2025

ICML 2025 (1)

• Improving Consistency Models with Generator-Augmented Flows — Spotlight. Thibaut Issenhuth, Sangchul Lee (AI/R Lab, KIST), Ludovic Dos Santos, Jean-Yves Franceschi, Chansoo Kim (AI/R Lab, KIST / Robot Dept., UST), Alain Rakotomamonjy

ICLR 2025 (2)

• Bringing NeRFs to the Latent Space: Inverse Graphics Autoencoder. Antoine Schnepf, Karim Kassab, Jean-Yves Franceschi, Laurent Caraffa (LASTIG / IGN / Université Gustave Eiffel), Flavian Vasile, Jérémie Mary, Andrew I. Comport (CNRS / I3S / Université Côte d’Azur), Valérie Gouet-Brunet (LASTIG / IGN / Université Gustave Eiffel)
• SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation. Song Duong (Liquid AI), Florian Le Bronnec (ISIR – Sorbonne Université), Alexandre Allauzen (LAMSADE / ESPCI), Vincent Guigue (UMR MIA-PS – AgroParisTech), Alberto Lumbreras, Laure Soulier (ISIR – Sorbonne Université), Patrick Gallinari (Sorbonne Université / Criteo AI Lab)

AISTATS 2025 (3)

• Prior-Dependent Allocations for Bayesian Fixed-Budget Best-Arm Identification in Structured Bandits. Nicolas Nguyen (University of Tübingen), Imad Aouali (CREST / Criteo AI Lab), András György (Google DeepMind), Claire Vernade (University of Tübingen)
• Bayesian Off-Policy Evaluation and Learning for Large Action Spaces. Imad Aouali (CREST / Criteo AI Lab), Victor-Emmanuel Brunel (CREST), David Rohde, Anna Korba (CREST)
• Distribution-Aware Mean Estimation under User-level Local Differential Privacy. Corentin Pla (CREST / Fairplay / Criteo AI Lab), Maxime Vono, Hugo Richard

TMLR 2025 (2)

• RefinedFields: Radiance Fields Refinement for Planar Scene Representations. Karim Kassab, Antoine Schnepf, Jean-Yves Franceschi, Laurent Caraffa (LASTIG / IGN / Université Gustave Eiffel), Jérémie Mary, Valérie Gouet-Brunet (LASTIG / IGN / Université Gustave Eiffel)
• Differentially Private Gradient Flow based on the Sliced Wasserstein Distance. Ilana Sebag, Muni Sreenivas Pydi (LAMSADE – PSL University), Jean-Yves Franceschi, Alain Rakotomamonjy, Mike Gartrell (Sigma Nova), Jamal Atif (LAMSADE – PSL University), Alexandre Allauzen (LAMSADE / ESPCI)