25 Hi! PARIS Papers Accepted at ICLR 2026
Hi! PARIS is proud to highlight the strong presence of its researchers at ICLR 2026, one of the world’s leading gatherings of professionals dedicated to the advancement of the branch of artificial intelligence called representation learning.
This year, 25 papers from Hi! PARIS affiliated researchers have been accepted, reflecting both the depth and diversity of the Center’s scientific contributions.
Their work spans both fundamental advances and applied research, covering topics such as efficient deep learning optimization, privacy in federated learning, policy gradient theory, uncertainty quantification, causal and decision-focused learning, graph and representation learning, optimal transport, and multimodal generative modeling.
Beyond theoretical progress, these papers also address concrete challenges in areas such as healthcare, cybersecurity, and data-driven decision-making. Several of them reflect collaborations across institutions and disciplines within the Hi! PARIS ecosystem.
This collective achievement once again highlights Hi! PARIS’ mission to advance AI research at the intersection of science, business, and society.
Congratulations to our researchers!
Here is a list of papers accepted at ICLR 2026 that include at least one author affiliated with Hi! PARIS:
| Title | Hi! PARIS authors | All Authors |
|---|---|---|
| Beyond Softmax and Entropy: Convergence Rates of Policy Gradients with $\boldsymbol{f}$-SoftArgmax Parameterization \& Coupled Regularization | Paul Mangold | Safwan Labbi, Daniil Tiapkin, Paul Mangold, Eric Moulines |
| Efficient Zero-shot Inpainting with Decoupled Diffusion Guidance | Alain Oliviero Durmus | Badr Moufad, Yazid Janati, Navid Bagheri Shouraki, Alain Oliviero Durmus, Thomas Hirtz, Eric Moulines, Jimmy Olsson |
| Online Decision-Focused Learning | Alain Oliviero Durmus | Aymeric Capitaine, Maxime Haddouche, Eric Moulines, Michael I. Jordan, Etienne Boursier, Alain Oliviero Durmus |
| Contextual Causal Bayesian Optimisation | Arnak Dalalyan | Vahan Arsenyan, Antoine Grosnit, Haitham Bou Ammar, Arnak S. Dalalyan |
| Protection against Source Inference Attacks in Federated Learning | Catuscia Palamidessi | Andreas Athanasiou, Kangsoo Jung, Catuscia Palamidessi |
| Winter Soldier: Backdooring Language Models at Pre-Training with Indirect Data Poisoning | El Mahdi El Mhamdi | Wassim Bouaziz, Mathurin VIDEAU, Nicolas Usunier, El‑Mahdi El‑Mhamdi |
| Efficient Resource-Constrained Training of Transformers via Subspace Optimization | Enzo Tartaglione, Van-Tam Nguyen | Le-Trung Nguyen, Enzo Tartaglione, Van-Tam Nguyen |
| INSTANT: Compressing Gradients and Activations for Resource-Efficient Training | Enzo Tartaglione, Van-tam Nguyen | Tuan-Kiet Doan, Trung-Hieu Tran, Enzo Tartaglione, Nikola Simidjievski, Van-Tam Nguyen |
| Study of Training Dynamics for Memory-Constrained Fine-Tuning | Enzo Tartaglione , Pavlo Mozharovskyi, Van-tam Nguyen | Aël Quélennec, Nour Hezbri, Pavlo Mozharovskyi, Van-Tam Nguyen, Enzo Tartaglione |
| Query-Level Uncertainty in Large Language Models | Fabian Suchanek, gaël varoquaux | Lihu Chen, Gerard de Melo, Fabian M. Suchanek, Gaël Varoquaux |
| Ensembling Pruned Attention Heads For Uncertainty-Aware Efficient Transformers | Gianni Franchi | Firas Gabetni, Giuseppe Curci, Andrea Pilzer, Subhankar Roy, Elisa Ricci, Gianni Franchi |
| Graph Representational Learning: When Does More Expressivity Hurt Generalization? | Johannes F. Lutzeyer | Sohir Maskey, Raffaele Paolino, Fabian Jogl, Gitta Kutyniok, Johannes F. Lutzeyer |
| A Spectral-Grassmann Wasserstein metric for operator representations of dynamical systems | Karim Lounici, Rémi Flamary | Thibaut Germain, Rémi Flamary, Vladimir R Kostic, Karim Lounici |
| Dynamic Reflections: Probing Video Representations with Text Alignment | Maks Ovsjanikov | Maks Ovsjanikov, Viorica Patraucean, Leonidas Guibas, Tyler Zhu, Tengda Han |
| Unified Brain Surface and Volume Registration | Maks Ovsjanikov | Mazdak Abulnaga, Andrew Hoopes, Malte Hoffmann, Robin Magnet, Maks Ovsjanikov, Lilla Zollei, John Guttag, Bruce Fischl, Adrian V Dalca |
| Completed Hyperparameter Transfer across Modules, Width, Depth, Batch and Duration | Marco Cuturi | Bruno Kacper Mlodozeniec; Pierre Ablin, Louis Béthune, Dan Busbridge, Michal Klein, Jason Ramapuram, marco cuturi |
| Flow Matching with Semidiscrete Couplings | Marco Cuturi | Alireza Mousavi-Hosseini, Stephen Y. Zhang, Michal Klein, marco cuturi |
| On the Impact of the Utility in Semivalue-based Data Valuation | Patrick Loiseau | Mélissa Tamine, Benjamin Heymann, Maxime Vono, Patrick Loiseau |
| Neural Optimal Transport Meets Multivariate Conformal Prediction | Rémi Flamary | Vladimir Kondratyev, Alexander Fishkov, Mahmoud Hegazy, Nikita Kotelevskii, Rémi Flamary, Maxim Panov, Eric Moulines |
| PSDNorm: Temporal Normalization for Deep Learning in Sleep Staging | Rémi Flamary | Theo Gnassounou, Antoine Collas, Rémi Flamary, Alexandre Gramfort |
| Soft-Di[M]O: Improving One-Step Discrete Image Generation with Soft Embeddings | Xi Wang, Stéphane Lathuilière, Vicky Kalogeiton | Yuanzhi Zhu, Xi Wang, Stéphane Lathuilière, Vicky Kalogeiton |
| Contextual Multi-Armed Bandits with Minimum Aggregated Revenue Constraints | Vianney Perchet | Ahmed Ben Yahmed, Hafedh El Ferchichi, Marc Abeille, Vianney Perchet |
| Learning in Prophet Inequalities with Noisy Observations | Vianney Perchet | Jung-hun Kim, Vianney Perchet |
| Pulp Motion: Framing-aware multimodal camera and human motion generation | Vicky Kalogeiton, Xi Wang | Robin Courant, Xi Wang, David Loiseaux, Marc Christie, Vicky Kalogeiton |
| Epistemic Uncertainty Quantification To Improve Decisions From Black-Box Models | Gaël Varoquaux | Sébastien Melo, Gaël Varoquaux, Marine Le Morvan |