Publications

Selected research contributions and a complete bibliography.

Selected contributions

A short entry point to recent work across representation learning, bilevel optimization, and reliable machine learning. The complete bibliography follows below.

PEIRA: Learning Predictive Encoders through Inter-View Regressor Alignment

NeurIPS 2026

Predictive self-supervision with an explicit objective linked to nonlinear canonical correlation.

Paper

BlockFormer: Transformer-based inference from genomic contact maps

NeurIPS 2026

Simulation-trained transformers locate centromeres from genomic contact maps across species.

Paper

EquiTabPFN: A Target-Permutation Equivariant Prior Fitted Network

NeurIPS 2025

Target-permutation equivariance removes class-order sensitivity and extrapolates to more classes.

Paper Code

Learning Theory for Kernel Bilevel Optimization

NeurIPS 2025

Finite-sample guarantees for generalization and gradient accuracy in kernel bilevel learning.

Paper

Functional Bilevel Optimization for Machine Learning

NeurIPS 2024 · Spotlight

Function-space bilevel optimization supports overparameterized neural inner models without strong convexity.

Paper Code

Amortized Implicit Differentiation for Stochastic Bilevel Optimization

ICLR 2022

Warm starts make implicit differentiation efficient without sacrificing optimal oracle complexity.

Paper Code

Peer-reviewed publications

  1. 2026
    Flow Matching Calibration for Simulation-Based Inference under Model Misspecification
    Pierre-Louis Ruhlmann and Michael Arbel and Florence Forbes and Pedro L. C. Rodrigues
    ICML 2026
  2. 2026
    Dual Perspectives on Non-Contrastive Self-Supervised Learning
    Jean Ponce and Basile Terver and Martial Hebert and Michael Arbel
    ICLR 2026
  3. 2026
    Beyond MMSE: Enhancing PnP Restoration with ProxiMAP
    Kenta Vert and Giacomo Meanti and Scott Pesme and Michael Arbel and Julien Mairal
    Advances in Neural Information Processing Systems (NeurIPS) 2026
  4. 2026
    BlockFormer : Transformer-based inference from genomic contact maps
    Eloïse Touron and Pedro L. C. Rodrigues and Julyan Arbel and Nelle Varoquaux and Michael Arbel
    Advances in Neural Information Processing Systems (NeurIPS) 2026
  5. 2026
    PEIRA: Learning Predictive Encoders through Inter-View Regressor Alignment
    Michael Arbel and Basile Terver and Jean Ponce
    Advances in Neural Information Processing Systems (NeurIPS) 2026
  6. 2026
    Non-Stationary Functional Bilevel Optimization
    Jason Bohne and Ieva Petrulionyte and Michael Arbel and Julien Mairal and Paweł Polak
    AISTATS 2026
  7. 2025
    Unsupervised Imaging Inverse Problems with Diffusion Distribution Matching
    Giacomo Meanti and Thomas Ryckeboer and Michael Arbel and Julien Mairal
    ICCV 2025
  8. 2025
    EquiTabPFN: A Target-Permutation Equivariant Prior Fitted Network
    Arbel, Michael and Salinas, David and Hutter, Frank
    Advances in Neural Information Processing Systems (NeurIPS) 2025
  9. 2025
    Learning Theory for Kernel Bilevel Optimization
    El Khoury, Fares and Pauwels, Edouard and Vaiter, Samuel and Arbel, Michael
    Advances in Neural Information Processing Systems (NeurIPS) 2025
  10. 2025
    LUDVIG: Learning-Free Uplifting of 2D Visual Features to Gaussian Splatting Scenes
    Marrie, Juliette and Ménégaux, Romain and Arbel, Michael and Larlus, Diane and Mairal, Julien
    IEEE/CVF International Conference on Computer Vision (ICCV) 2025
  11. 2025
    MAP Estimation with Denoisers: Convergence Rates and Guarantees
    Scott Pesme and Giacomo Meanti and Michael Arbel and Julien Mairal
    Advances in Neural Information Processing Systems (NeurIPS) 2025
  12. 2024
    Functional Bilevel Optimization for Machine Learning
    Petrulionyte, Ieva and Mairal, Julien and Arbel, Michael
    Advances in Neural Information Processing Systems (NeurIPS) 2024
  13. 2024
    MLXP: A framework for conducting replicable experiments in Python
    Arbel, Michael and Zouaoui, Alexander
    Proceedings of the 2nd ACM Conference on Reproducibility and Replicability
  14. 2024
    On Good Practices for Task-Specific Distillation of Large Pretrained Visual Models
    Marrie, Juliette and Arbel, Michael and Mairal, Julien and Larlus, Diane
    Transactions on Machine Learning Research Journal
  15. 2023
    Rethinking Gauss-Newton for learning over-parameterized models
    Arbel, Michael and Menegaux, Romain and Wolinski, Pierre
    Advances in Neural Information Processing Systems (NeurIPS) 2023
  16. 2023
    SLACK: Stable Learning of Augmentations with Cold-start and KL regularization
    Marrie, Juliette and Arbel, Michael and Larlus, Diane and Mairal, Julien
    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition
  17. 2022
    Amortized implicit differentiation for stochastic bilevel optimization
    Arbel, Michael and Mairal, Julien
    International Conference on Learning Representations (ICLR) 2022
  18. 2022
    Continual repeated annealed flow transport Monte Carlo
    Matthews, Alex and Arbel, Michael and Rezende, Danilo Jimenez and Doucet, Arnaud
    International Conference on Machine Learning
  19. 2022
    Non-Convex Bilevel Games with Critical Point Selection Maps
    Arbel, Michael and Mairal, Julien
    Advances in Neural Information Processing Systems (NeurIPS) 2022
  20. 2022
    Towards an Understanding of Default Policies in Multitask Policy Optimization
    Moskovitz, Ted and Arbel, Michael and Parker-Holder, Jack and Pacchiano, Aldo
    International Conference on Artificial Intelligence and Statistics (AISTATS) 2022
  21. 2021
    Annealed Flow Transport Monte Carlo
    Arbel, Michael and Matthews, Alexander GDG and Doucet, Arnaud
    International Conference on Machine Learning (ICML) 2021
  22. 2021
    Efficient wasserstein natural gradients for reinforcement learning
    Moskovitz, Ted and Arbel, Michael and Huszar, Ferenc and Gretton, Arthur
    International Conference on Learning Representations (ICLR) 2021
  23. 2021
    Generalized Energy Based Models
    Arbel, Michael and Zhou, Liang and Gretton, Arthur
    International Conference on Learning Representations (ICLR) 2021
  24. 2021
    KALE Flow: A Relaxed KL Gradient Flow for Probabilities with Disjoint Support
    Glaser, Pierre and Arbel, Michael and Gretton, Arthur
    Advances in Neural Information Processing Systems (NeurIPS) 2021
  25. 2021
    Tactical Optimism and Pessimism for Deep Reinforcement Learning
    Moskovitz, Ted and Parker-Holder, Jack and Pacchiano, Aldo and Arbel, Michael and Jordan, Michael I.
    Advances in Neural Information Processing Systems (NeurIPS) 2021
  26. 2021
    The Unreasonable Effectiveness of Patches in Deep Convolutional Kernels Methods
    Thiry, Louis and Arbel, Michael and Belilovsky, Eugene and Oyallon, Edouard
    International Conference on Learning Representations (ICLR) 2021
  27. 2020
    A non-asymptotic analysis for Stein variational gradient descent
    Korba, Anna and Salim, Adil and Arbel, Michael and Luise, Giulia and Gretton, Arthur
    Advances in Neural Information Processing Systems (NeurIPS) 2020
  28. 2020
    Kernelized Wasserstein Natural Gradient
    Arbel, Michael and Gretton, Arthur and Li, Wuchen and Montúfar, Guido
    International Conference on Learning Representations (ICLR) 2020
  29. 2020
    Synchronizing probability measures on rotations via optimal transport
    Birdal, Tolga and Arbel, Michael and Simsekli, Umut and Guibas, Leonidas J
    IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2020
  30. 2019
    Maximum mean discrepancy gradient flow
    Arbel, Michael and Korba, Anna and Salim, Adil and Gretton, Arthur
    Advances in Neural Information Processing Systems (NeurIPS) 2019
  31. 2018
    Demystifying mmd gans
    Bińkowski, Mikołaj and Sutherland, Dougal J and Arbel, Michael and Gretton, Arthur
    International Conference on Learning Representations (ICLR) 2018
  32. 2018
    Efficient and principled score estimation with Nyström kernel exponential families
    Sutherland, Dougal J and Strathmann, Heiko and Arbel, Michael and Gretton, Arthur
    International Conference on Artificial Intelligence and Statistics (AISTATS) 2018
  33. 2018
    Kernel conditional exponential family
    Arbel, Michael and Gretton, Arthur
    International Conference on Artificial Intelligence and Statistics (AISTATS) 2018
  34. 2018
    On gradient regularizers for MMD GANs
    Arbel, Michael and Sutherland, Danica J and Bińkowski, Mikołaj and Gretton, Arthur
    Advances in Neural Information Processing Systems (NeurIPS) 2018

Preprints

  1. 2026
    Functional Bilevel Optimization for Predictive Fairness
    Ieva Petrulionyte and Julien Mairal and Michael Arbel
    arXiv preprint arXiv:2607.05098
  2. 2026
    Semiparametric Efficient Bilevel Gradient Estimation
    El Khoury, Fares and Zenati, Houssam and Kallus, Nathan and Arbel, Michael and Bibaut, Aurélien
    arXiv preprint arXiv:2605.21341
  3. 2022
    Maximum Likelihood Learning of Unnormalized Models for Simulation-Based Inference
    Glaser, Pierre and Arbel, Michael and Hromadka, Samo and Doucet, Arnaud and Gretton, Arthur
    arXiv preprint arXiv:2210.14756
  4. 2020
    Estimating barycenters of measures in high dimensions
    Cohen, Samuel and Arbel, Michael and Deisenroth, Marc Peter
    arXiv preprint arXiv:2007.07105

Other research outputs

  1. 2025
    Simulation-based inference of yeast centromeres
    Eloïse Touron and Pedro L. C. Rodrigues and Julyan Arbel and Nelle Varoquaux and Michael Arbel
    Advances in Neural Information Processing Systems (NeurIPS) 2025, 3rd Workshop on Imageomic
  2. 2024
    Data augmentation for training neural networks
    Marrie, Juliette and Arbel, Michael and Mairal, Julien and Larlus, Diane
    US Patent Application 18/634,466
  3. 2023
    Improving Deep Ensembles without Communication
    Pitas, Konstantinos and Arbel, Michael and Arbel, Julyan
    Workshop on Advancing Neural Network Training (WANT@ NeurIPS 2023)
  4. 2022
    Fair synthetic data does not necessarily lead to fair models
    Eitan, Yam and Cavaglione, Nathan and Arbel, Michael and Cohen, Samuel
    NeurIPS 2022 Workshop on Synthetic Data for Empowering ML Research
  5. 2021
    Methods for optimization and regularization of Generative Models
    Arbel, Michael
    UCL (University College London)