Zobrazeno 1 - 10
of 68
pro vyhledávání: '"Berar , Maxime"'
Autor:
Piquenot, Jason, Bérar, Maxime, Héroux, Pierre, Ramel, Jean-Yves, Raveaux, Romain, Adam, Sébastien
This paper presents Grammar Reinforcement Learning (GRL), a reinforcement learning algorithm that uses Monte Carlo Tree Search (MCTS) and a transformer architecture that models a Pushdown Automaton (PDA) within a context-free grammar (CFG) framework.
Externí odkaz:
http://arxiv.org/abs/2410.01661
Gaussian smoothed sliced Wasserstein distance has been recently introduced for comparing probability distributions, while preserving privacy on the data. It has been shown that it provides performances similar to its non-smoothed (non-private) counte
Externí odkaz:
http://arxiv.org/abs/2404.03273
Contrastive learning has demonstrated great effectiveness in representation learning especially for image classification tasks. However, there is still a shortage in the studies targeting regression tasks, and more specifically applications on hypers
Externí odkaz:
http://arxiv.org/abs/2403.17014
Autor:
Piquenot, Jason, Moscatelli, Aldo, Bérar, Maxime, Héroux, Pierre, raveaux, Romain, Ramel, Jean-Yves, Adam, Sébastien
This paper introduces a framework for formally establishing a connection between a portion of an algebraic language and a Graph Neural Network (GNN). The framework leverages Context-Free Grammars (CFG) to organize algebraic operations into generative
Externí odkaz:
http://arxiv.org/abs/2303.01590
Categorical data are present in key areas such as health or supply chain, and this data require specific treatment. In order to apply recent machine learning models on such data, encoding is needed. In order to build interpretable models, one-hot enc
Externí odkaz:
http://arxiv.org/abs/2209.03771
Gaussian smoothed sliced Wasserstein distance has been recently introduced for comparing probability distributions, while preserving privacy on the data. It has been shown, in applications such as domain adaptation, to provide performances similar to
Externí odkaz:
http://arxiv.org/abs/2110.10524
Publikováno v:
In Pattern Recognition Letters August 2024 184:14-20
Optimal Transport (OT) metrics allow for defining discrepancies between two probability measures. Wasserstein distance is for longer the celebrated OT-distance frequently-used in the literature, which seeks probability distributions to be supported o
Externí odkaz:
http://arxiv.org/abs/2106.02542
Autor:
Rakotomamonjy, Alain, Flamary, Rémi, Gasso, Gilles, Alaya, Mokhtar Z., Berar, Maxime, Courty, Nicolas
We address the problem of unsupervised domain adaptation under the setting of generalized target shift (joint class-conditional and label shifts). For this framework, we theoretically show that, for good generalization, it is necessary to learn a lat
Externí odkaz:
http://arxiv.org/abs/2006.08161
We propose a novel approach for comparing distributions whose supports do not necessarily lie on the same metric space. Unlike Gromov-Wasserstein (GW) distance which compares pairwise distances of elements from each distribution, we consider a method
Externí odkaz:
http://arxiv.org/abs/2002.08314