publications

2026

1. CVPR

   ReBaPL: Repulsive Bayesian Prompt Learning

   Yassir Bendou, Omar Ezzahir, Eduardo Fernandes Montesuma, and 3 more authors

   In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026

   Abs arXiv Bib PDF

   Conventional prompt learning methods are prone to overfitting and can struggle with out-of-distribution generalization. We frame prompt optimization as a Bayesian inference problem and propose ReBaPL, which integrates a cyclical step-size schedule with a stochastic gradient Hamiltonian Monte Carlo (SGHMC) algorithm and a repulsive force derived from probability metrics (Maximum Mean Discrepancy, Wasserstein) to diversify exploration of the prompt posterior and prevent premature collapse to a single mode.

   @inproceedings{bendou2026rebapl,
     title = {ReBaPL: Repulsive Bayesian Prompt Learning},
     author = {Bendou, Yassir and Ezzahir, Omar and Fernandes Montesuma, Eduardo and Mahuas, Gabriel and Shevchenko, Victoria and Gartrell, Mike},
     booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
     year = {2026},
   }

2. arXiv

   Wasserstein Gradient Flows for Scalable and Regularized Barycenter Computation

   Eduardo Fernandes Montesuma, Yassir Bendou, and Mike Gartrell

   2026

   Under review

   Abs arXiv Bib PDF

   We present a method for computing Wasserstein barycenters via gradient flows in probability measure space. The approach yields a scalable algorithm that (i) relies on mini-batch optimal transport, (ii) accepts modular regularization through task-aware functions, and (iii) admits supervised information in the cost function. Experiments on domain adaptation tasks across computer vision, neuroscience, and chemical engineering show that incorporating label information improves over existing barycenter approaches.

   @misc{montesuma2026wgf,
     title = {Wasserstein Gradient Flows for Scalable and Regularized Barycenter Computation},
     author = {Fernandes Montesuma, Eduardo and Bendou, Yassir and Gartrell, Mike},
     year = {2026},
     note = {Under review},
   }

3. arXiv

   Structure-Preserving Multi-View Embedding Using Gromov-Wasserstein Optimal Transport

   Rafael Pereira Eufrazio, Eduardo Fernandes Montesuma, and Charles Casimiro Cavalcante

   2026

   Under review at Signal Processing (Elsevier)

   arXiv Bib PDF

   @misc{eufrazio2026structure,
     title = {Structure-Preserving Multi-View Embedding Using Gromov-Wasserstein Optimal Transport},
     author = {Eufrazio, Rafael Pereira and Fernandes Montesuma, Eduardo and Cavalcante, Charles Casimiro},
     year = {2026},
     note = {Under review at Signal Processing (Elsevier)},
   }

4. arXiv

   Gromov-Wasserstein Methods for Multi-View Relational Embedding and Clustering

   Rafael Pereira Eufrazio, Eduardo Fernandes Montesuma, and Charles Casimiro Cavalcante

   2026

   Under review at SBrT 2026

   arXiv Bib PDF

   @misc{eufrazio2026gromov,
     title = {Gromov-Wasserstein Methods for Multi-View Relational Embedding and Clustering},
     author = {Eufrazio, Rafael Pereira and Fernandes Montesuma, Eduardo and Cavalcante, Charles Casimiro},
     year = {2026},
     note = {Under review at SBrT 2026},
   }

2025

1. TMLR

   Optimal Transport for Domain Adaptation through Gaussian Mixture Models

   Eduardo Fernandes Montesuma, Fred Maurice Ngolè Mboula, and Antoine Souloumiac

   Transactions on Machine Learning Research, 2025

   arXiv Bib PDF

   @article{montesuma2025optimal,
     title = {Optimal Transport for Domain Adaptation through Gaussian Mixture Models},
     author = {Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e} and Souloumiac, Antoine},
     journal = {Transactions on Machine Learning Research},
     issn = {2835-8856},
     year = {2025},
     url = {https://openreview.net/forum?id=DCAeXwLenB},
   }

2. TMLR

   Unsupervised Anomaly Detection through Mass Repulsing Optimal Transport

   Eduardo Fernandes Montesuma, Adel El Habazi, and Fred Maurice Ngolè Mboula

   Transactions on Machine Learning Research, 2025

   arXiv Bib PDF

   @article{montesuma2025unsupervised,
     title = {Unsupervised Anomaly Detection through Mass Repulsing Optimal Transport},
     author = {Fernandes Montesuma, Eduardo and El Habazi, Adel and Mboula, Fred Maurice Ngol{\`e}},
     journal = {Transactions on Machine Learning Research},
     issn = {2835-8856},
     year = {2025},
     url = {https://openreview.net/forum?id=PPGJ3EvENv},
   }

3. GSI

   KD\textsuperscript2M: A Unifying Framework for Feature Knowledge Distillation

   Eduardo Fernandes Montesuma

   In Geometric Science of Information (GSI), 2025

   arXiv Bib PDF

   @inproceedings{montesuma2025kd2m,
     title = {KD\textsuperscript{2}M: A Unifying Framework for Feature Knowledge Distillation},
     author = {Fernandes Montesuma, Eduardo},
     booktitle = {Geometric Science of Information (GSI)},
     year = {2025},
   }

4. GSI

   A Dimensionality Reduction Technique Based on the Gromov-Wasserstein Distance

   Rafael Pereira Eufrazio, Eduardo Fernandes Montesuma, and Charles Casimiro Cavalcante

   In Geometric Science of Information (GSI), 2025

   arXiv Bib PDF

   @inproceedings{eufrazio2025dimensionality,
     title = {A Dimensionality Reduction Technique Based on the Gromov-Wasserstein Distance},
     author = {Eufrazio, Rafael Pereira and Fernandes Montesuma, Eduardo and Cavalcante, Charles Casimiro},
     booktitle = {Geometric Science of Information (GSI)},
     year = {2025},
   }

5. ICASSP

   Decentralized Federated Dataset Dictionary Learning for Multi-Source Domain Adaptation

   Rebecca Clain, Eduardo Fernandes Montesuma, and Fred Maurice Ngolè Mboula

   In ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025

   arXiv Bib PDF

   @inproceedings{clain2025decentralized,
     title = {Decentralized Federated Dataset Dictionary Learning for Multi-Source Domain Adaptation},
     author = {Clain, Rebecca and Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e}},
     booktitle = {ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
     pages = {1--5},
     year = {2025},
     organization = {IEEE},
   }

2024

1. TPAMI

   Recent Advances in Optimal Transport for Machine Learning

   Eduardo Fernandes Montesuma, Fred Maurice Ngolè Mboula, and Antoine Souloumiac

   IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

   Abs arXiv Bib PDF

   Optimal Transport (OT) is a probabilistic framework for comparing and manipulating probability distributions. We survey the contributions of OT to machine learning over 2012-2023 across supervised, unsupervised, transfer, and reinforcement learning, including computational OT, partial, unbalanced, Gromov, and Neural OT and their interplay with ML practice.

   @article{montesuma2024recent,
     title = {Recent Advances in Optimal Transport for Machine Learning},
     author = {Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e} and Souloumiac, Antoine},
     journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
     year = {2024},
     publisher = {IEEE},
   }

2. ECML-PKDD

   Lighter, Better, Faster Multi-Source Domain Adaptation with Gaussian Mixture Models and Optimal Transport

   Eduardo Fernandes Montesuma, Fred Maurice Ngolè Mboula, and Antoine Souloumiac

   In Joint European Conference on Machine Learning and Knowledge Discovery in Databases (ECML-PKDD), 2024

   arXiv Bib PDF Code

   @inproceedings{montesuma2024lighter,
     title = {Lighter, Better, Faster Multi-Source Domain Adaptation with Gaussian Mixture Models and Optimal Transport},
     author = {Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e} and Souloumiac, Antoine},
     booktitle = {Joint European Conference on Machine Learning and Knowledge Discovery in Databases (ECML-PKDD)},
     pages = {21--38},
     year = {2024},
     organization = {Springer},
   }

3. ICASSP

   Multi-Source Domain Adaptation Meets Dataset Distillation through Dataset Dictionary Learning

   Eduardo Fernandes Montesuma, Fred Maurice Ngolè Mboula, and Antoine Souloumiac

   In ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024

   arXiv Bib PDF

   @inproceedings{montesuma2024multi,
     title = {Multi-Source Domain Adaptation Meets Dataset Distillation through Dataset Dictionary Learning},
     author = {Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e} and Souloumiac, Antoine},
     booktitle = {ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
     pages = {5620--5624},
     year = {2024},
     organization = {IEEE},
   }

4. MLSP

   Online Multi-Source Domain Adaptation Through Gaussian Mixtures and Dataset Dictionary Learning

   Eduardo Fernandes Montesuma, Stevan Le Stanc, and Fred Maurice Ngolè Mboula

   In 2024 IEEE 34th International Workshop on Machine Learning for Signal Processing (MLSP), 2024

   arXiv Bib PDF

   @inproceedings{montesuma2024online,
     title = {Online Multi-Source Domain Adaptation Through Gaussian Mixtures and Dataset Dictionary Learning},
     author = {Fernandes Montesuma, Eduardo and Le Stanc, Stevan and Mboula, Fred Maurice Ngol{\`e}},
     booktitle = {2024 IEEE 34th International Workshop on Machine Learning for Signal Processing (MLSP)},
     pages = {1--6},
     year = {2024},
     organization = {IEEE},
   }

5. arXiv

   Dataset Dictionary Learning in a Wasserstein Space for Federated Domain Adaptation

   Eduardo Fernandes Montesuma, Fabiola Espinoza Castellon, Fred Maurice Ngolè Mboula, and 3 more authors

   2024

   arXiv preprint

   arXiv Bib PDF

   @misc{montesuma2024dataset,
     title = {Dataset Dictionary Learning in a Wasserstein Space for Federated Domain Adaptation},
     author = {Fernandes Montesuma, Eduardo and Castellon, Fabiola Espinoza and Mboula, Fred Maurice Ngol{\`e} and Mayoue, Aur{\'e}lien and Souloumiac, Antoine and Gouy-Pailler, C{\'e}dric},
     year = {2024},
     note = {arXiv preprint},
   }

6. ICASSP

   Federated Dataset Dictionary Learning for Multi-Source Domain Adaptation

   Fabiola Espinoza Castellon, Eduardo Fernandes Montesuma, Fred Maurice Ngolè Mboula, and 3 more authors

   In ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024

   arXiv Bib PDF

   @inproceedings{castellon2024federated,
     title = {Federated Dataset Dictionary Learning for Multi-Source Domain Adaptation},
     author = {Espinoza Castellon, Fabiola and Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e} and Mayoue, Aur{\'e}lien and Souloumiac, Antoine and Gouy-Pailler, C{\'e}dric},
     booktitle = {ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
     pages = {5610--5614},
     year = {2024},
     organization = {IEEE},
   }

7. PhD

   Multi-Source Domain Adaptation through Wasserstein Barycenters

   Eduardo Fernandes Montesuma

   Université Paris-Saclay, Sep 2024

   Bib PDF

   @phdthesis{fernandesmontesuma2024thesis,
     title = {Multi-Source Domain Adaptation through Wasserstein Barycenters},
     author = {Fernandes Montesuma, Eduardo},
     year = {2024},
     month = sep,
     school = {Universit{\'e} Paris-Saclay},
     number = {2024UPASG045},
     url = {https://theses.hal.science/tel-04743619},
   }

2023

1. ECAI

   Multi-Source Domain Adaptation through Dataset Dictionary Learning in Wasserstein Space

   Eduardo Fernandes Montesuma, Fred Maurice Ngolè Mboula, and Antoine Souloumiac

   In 26th European Conference on Artificial Intelligence (ECAI), Sep 2023

   arXiv Bib PDF Code

   @inproceedings{montesuma2023multi,
     title = {Multi-Source Domain Adaptation through Dataset Dictionary Learning in Wasserstein Space},
     author = {Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e} and Souloumiac, Antoine},
     booktitle = {26th European Conference on Artificial Intelligence (ECAI)},
     pages = {1739--1746},
     year = {2023},
     publisher = {IOS Press},
   }

2. ML4CCE

   Benchmarking Domain Adaptation for Chemical Processes on the Tennessee Eastman Process

   Eduardo Fernandes Montesuma, Michela Mulas, Fred Maurice Ngolè Mboula, and 2 more authors

   Sep 2023

   Machine Learning for Chemistry and Chemical Engineering (ML4CCE) Workshop at ECML-PKDD 2024

   arXiv Bib PDF

   @misc{montesuma2023benchmarking,
     title = {Benchmarking Domain Adaptation for Chemical Processes on the Tennessee Eastman Process},
     author = {Fernandes Montesuma, Eduardo and Mulas, Michela and Mboula, Fred Maurice Ngol{\`e} and Corona, Francesco and Souloumiac, Antoine},
     year = {2023},
     note = {Machine Learning for Chemistry and Chemical Engineering (ML4CCE) Workshop at ECML-PKDD 2024},
   }

2022

1. IFAC

   Cross-Domain Fault Diagnosis through Optimal Transport for a CSTR Process

   Eduardo Fernandes Montesuma, Michela Mulas, Francesco Corona, and 1 more author

   IFAC-PapersOnLine, Sep 2022

   Bib PDF

   @article{montesuma2022cross,
     title = {Cross-Domain Fault Diagnosis through Optimal Transport for a CSTR Process},
     author = {Fernandes Montesuma, Eduardo and Mulas, Michela and Corona, Francesco and Mboula, Fred Maurice Ngol{\`e}},
     journal = {IFAC-PapersOnLine},
     volume = {55},
     number = {7},
     pages = {946--951},
     year = {2022},
     publisher = {Elsevier},
   }

2021

1. CVPR

   Wasserstein Barycenter for Multi-Source Domain Adaptation

   Eduardo Fernandes Montesuma and Fred Maurice Ngolè Mboula

   In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Sep 2021

   Abs Bib PDF

   Multi-source domain adaptation is a key technique that allows a model to be trained on data coming from various probability distributions. We propose a method for constructing an intermediate domain between sources and target domain via the barycenter on Wasserstein spaces (WBT). After aggregating sources, they are transported to the target domain using standard OT for Domain Adaptation. We evaluate on object recognition, face recognition, music genre recognition, and music-speech discrimination.

   @inproceedings{montesuma2021wasserstein,
     title = {Wasserstein Barycenter for Multi-Source Domain Adaptation},
     author = {Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e}},
     booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
     pages = {16785--16793},
     year = {2021},
   }

2. ICASSP

   Wasserstein Barycenter Transport for Acoustic Adaptation

   Eduardo Fernandes Montesuma and Fred Maurice Ngolè Mboula

   In ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2021

   Bib

   @inproceedings{montesuma2021acoustic,
     title = {Wasserstein Barycenter Transport for Acoustic Adaptation},
     author = {Fernandes Montesuma, Eduardo and Mboula, Fred Maurice Ngol{\`e}},
     booktitle = {ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
     pages = {3405--3409},
     year = {2021},
     organization = {IEEE},
     url = {https://ieeexplore.ieee.org/abstract/document/9414199}
   }

3. STIL

   An Empirical Study of Information Retrieval and Machine Reading Comprehension Algorithms for an Online Education Platform

   Eduardo Fernandes Montesuma, Lucas C Carneiro, Adson R P Damasceno, and 4 more authors

   In Simpósio Brasileiro de Tecnologia da Informação e da Linguagem Humana (STIL), Sep 2021

   Bib PDF

   @inproceedings{montesuma2021empirical,
     title = {An Empirical Study of Information Retrieval and Machine Reading Comprehension Algorithms for an Online Education Platform},
     author = {Fernandes Montesuma, Eduardo and Carneiro, Lucas C and Damasceno, Adson R P and de Sampaio, Joao Victor F T and F{\'e}rrer Filho, Romulo F and Maia, Paulo Henrique M and Oliveira, Francisco C M B},
     booktitle = {Simp{\'o}sio Brasileiro de Tecnologia da Informa{\c{c}}{\~a}o e da Linguagem Humana (STIL)},
     pages = {217--226},
     year = {2021},
     organization = {SBC},
   }

2020

1. ICASSP

   OpenDenoising: An Extensible Benchmark for Building Comparative Studies of Image Denoisers

   Florian Lemarchand, Eduardo Fernandes Montesuma, Maxime Pelcat, and 1 more author

   In ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2020

   arXiv Bib PDF

   @inproceedings{lemarchand2020opendenoising,
     title = {OpenDenoising: An Extensible Benchmark for Building Comparative Studies of Image Denoisers},
     author = {Lemarchand, Florian and Fernandes Montesuma, Eduardo and Pelcat, Maxime and Nogues, Erwan},
     booktitle = {ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
     pages = {2648--2652},
     year = {2020},
     organization = {IEEE},
   }

2017

1. SIIM

   Avaliação de Algoritmos de Classificação de Padrões na Detecção de Câncer do Colo do Útero

   Eduardo Fernandes Montesuma, Levi Pio Souza Arrais Alencar, and Guilherme de Alencar Barreto

   In VIII Simpósio de Instrumentação e Imagens Médicas (SIIM), Sep 2017

   Bib

   @inproceedings{montesuma2017avaliacao,
     title = {Avalia{\c{c}}{\~a}o de Algoritmos de Classifica{\c{c}}{\~a}o de Padr{\~o}es na Detec{\c{c}}{\~a}o de C{\^a}ncer do Colo do {\'U}tero},
     author = {Fernandes Montesuma, Eduardo and Alencar, Levi Pio Souza Arrais and Barreto, Guilherme de Alencar},
     booktitle = {VIII Simp{\'o}sio de Instrumenta{\c{c}}{\~a}o e Imagens M{\'e}dicas (SIIM)},
     year = {2017}
   }