Aurélien Decelle
YOU?
Author Swipe
View article: On the role of non-linear latent features in bipartite generative neural networks
On the role of non-linear latent features in bipartite generative neural networks Open
We investigate the phase diagram and memory retrieval capabilities of Restricted Boltzmann Machines (RBMs), an archetypal model of bipartite energy-based neural networks, as a function of the prior distribution imposed on their hidden unit…
View article: Inferring Higher-Order Couplings with Neural Networks
Inferring Higher-Order Couplings with Neural Networks Open
Maximum entropy methods, rooted in the inverse Ising/Potts problem from statistical physics, are widely used to model pairwise interactions in complex systems across disciplines such as bioinformatics and neuroscience. While successful, th…
View article: PRIVET: Privacy Metric Based on Extreme Value Theory
PRIVET: Privacy Metric Based on Extreme Value Theory Open
Deep generative models are often trained on sensitive data, such as genetic sequences, health data, or more broadly, any copyrighted, licensed or protected content. This raises critical concerns around privacy-preserving synthetic data, an…
View article: Cascade of phase transitions in the training of energy-based models<sup>*</sup>
Cascade of phase transitions in the training of energy-based models<sup>*</sup> Open
In this paper, we investigate the feature encoding process in a prototypical energy-based generative model, the restricted Boltzmann machine (RBM). We start with an analytical investigation using simplified architectures and data structure…
View article: Fast training and sampling of Restricted Boltzmann Machines
Fast training and sampling of Restricted Boltzmann Machines Open
15 figures, 31 pages
View article: A theoretical framework for overfitting in energy-based modeling
A theoretical framework for overfitting in energy-based modeling Open
We investigate the impact of limited data on training pairwise energy-based models for inverse problems aimed at identifying interaction networks. Utilizing the Gaussian model as testbed, we dissect training trajectories across the eigenba…
View article: Fast and Functional Structured Data Generators Rooted in Out-of-Equilibrium Physics
Fast and Functional Structured Data Generators Rooted in Out-of-Equilibrium Physics Open
In this study, we address the challenge of using energy-based models to produce high-quality, label-specific data in complex structured datasets, such as population genetics, RNA or protein sequences data. Traditional training methods enco…
View article: Fast training and sampling of Restricted Boltzmann Machines
Fast training and sampling of Restricted Boltzmann Machines Open
Restricted Boltzmann Machines (RBMs) are powerful tools for modeling complex systems and extracting insights from data, but their training is hindered by the slow mixing of Markov Chain Monte Carlo (MCMC) processes, especially with highly …
View article: Inferring effective couplings with restricted Boltzmann machines
Inferring effective couplings with restricted Boltzmann machines Open
Generative models offer a direct way of modeling complex data. Energy-based models attempt to encode the statistical correlations observed in the data at the level of the Boltzmann weight associated with an energy function in the form of a…
View article: Predicting large scale cosmological structure evolution with GAN-based autoencoders
Predicting large scale cosmological structure evolution with GAN-based autoencoders Open
Cosmological simulations play a key role in the prediction and understanding of large scale structure formation from initial conditions. We make use of GAN-based Autoencoders (AEs) in an attempt to predict structure evolution within simula…
View article: Cascade of phase transitions in the training of energy-based models
Cascade of phase transitions in the training of energy-based models Open
In this paper, we investigate the feature encoding process in a prototypical energy-based generative model, the Restricted Boltzmann Machine (RBM). We start with an analytical investigation using simplified architectures and data structure…
View article: Report on 2309.02292v2
Report on 2309.02292v2 Open
Generative models offer a direct way to model complex data.Among them, energy-based models provide us with a neural network model that aims to accurately reproduce all statistical correlations observed in the data at the level of the Boltz…
View article: Deep convolutional and conditional neural networks for large-scale genomic data generation
Deep convolutional and conditional neural networks for large-scale genomic data generation Open
Applications of generative models for genomic data have gained significant momentum in the past few years, with scopes ranging from data characterization to generation of genomic segments and functional sequences. In our previous study, we…
View article: Report on 2309.02292v2
Report on 2309.02292v2 Open
Generative models offer a direct way to model complex data.Among them, energy-based models provide us with a neural network model that aims to accurately reproduce all statistical correlations observed in the data at the level of the Boltz…
View article: The Copycat Perceptron: Smashing Barriers Through Collective Learning
The Copycat Perceptron: Smashing Barriers Through Collective Learning Open
We characterize the equilibrium properties of a model of $y$ coupled binary perceptrons in the teacher-student scenario, subject to a suitable cost function, with an explicit ferromagnetic coupling proportional to the Hamming distance betw…
View article: The Mighty Force: Statistical Inference and High-Dimensional Statistics
The Mighty Force: Statistical Inference and High-Dimensional Statistics Open
This is a review to appear as a contribution to the edited volume "Spin Glass Theory & Far Beyond - Replica Symmetry Breaking after 40 Years", World Scientific. It showcases a selection of contributions from the spin glass community at lar…
View article: Fast and Functional structured data generator
Fast and Functional structured data generator Open
International audience
View article: Fast and Functional Structured Data Generators Rooted in Out-of-Equilibrium Physics
Fast and Functional Structured Data Generators Rooted in Out-of-Equilibrium Physics Open
In this study, we address the challenge of using energy-based models to produce high-quality, label-specific data in complex structured datasets, such as population genetics, RNA or protein sequences data. Traditional training methods enco…
View article: Unsupervised hierarchical clustering using the learning dynamics of restricted Boltzmann machines
Unsupervised hierarchical clustering using the learning dynamics of restricted Boltzmann machines Open
Data sets in the real world are often complex and to some degree hierarchical, with groups and subgroups of data sharing common characteristics at different levels of abstraction. Understanding and uncovering the hidden structure of these …
View article: Cosmology with cosmic web environments
Cosmology with cosmic web environments Open
Degeneracies among parameters of the cosmological model are known to drastically limit the information contained in the matter distribution. In the first paper of this series, we show that the cosmic web environments, namely the voids, wal…
View article: Thermodynamics of bidirectional associative memories
Thermodynamics of bidirectional associative memories Open
In this paper we investigate the equilibrium properties of bidirectional associative memories (BAMs). Introduced by Kosko in 1988 as a generalization of the Hopfield model to a bipartite structure, the simplest architecture is defined by t…
View article: Learning a restricted Boltzmann machine using biased Monte Carlo sampling
Learning a restricted Boltzmann machine using biased Monte Carlo sampling Open
Restricted Boltzmann Machines are simple and powerful generative models that can encode any complex dataset. Despite all their advantages, in practice the trainings are often unstable and it is difficult to assess their quality because the…
View article: Deep convolutional and conditional neural networks for large-scale genomic data generation
Deep convolutional and conditional neural networks for large-scale genomic data generation Open
Applications of generative models for genomic data have gained significant momentum in the past few years, with scopes ranging from data characterization to generation of genomic segments and functional sequences. In our previous study, we…
View article: Unsupervised hierarchical clustering using the learning dynamics of RBMs
Unsupervised hierarchical clustering using the learning dynamics of RBMs Open
Datasets in the real world are often complex and to some degree hierarchical, with groups and sub-groups of data sharing common characteristics at different levels of abstraction. Understanding and uncovering the hidden structure of these …
View article: Explaining the effects of non-convergent sampling in the training of Energy-Based Models
Explaining the effects of non-convergent sampling in the training of Energy-Based Models Open
In this paper, we quantify the impact of using non-convergent Markov chains to train Energy-Based models (EBMs). In particular, we show analytically that EBMs trained with non-persistent short runs to estimate the gradient can perfectly re…
View article: Cosmology with cosmic web environments II. Redshift-space auto and cross power spectra
Cosmology with cosmic web environments II. Redshift-space auto and cross power spectra Open
Degeneracies among parameters of the cosmological model are known to drastically limit the information contained in the matter distribution. In the first paper of this series, we shown that the cosmic web environments; namely the voids, wa…
View article: Thermodynamics of bidirectional associative memories
Thermodynamics of bidirectional associative memories Open
In this paper we investigate the equilibrium properties of bidirectional\nassociative memories (BAMs). Introduced by Kosko in 1988 as a generalization of\nthe Hopfield model to a bipartite structure, the simplest architecture is\ndefined b…
View article: Thermodynamics of bidirectional associative memories
Thermodynamics of bidirectional associative memories Open
In this paper we investigate the equilibrium properties of bidirectional associative memories (BAMs). Introduced by Kosko in 1988 as a generalization of the Hopfield model to a bipartite structure, the simplest architecture is defined by t…
View article: Equilibrium and non-equilibrium regimes in the learning of restricted Boltzmann machines*
Equilibrium and non-equilibrium regimes in the learning of restricted Boltzmann machines* Open
Training restricted Boltzmann machines (RBMs) have been challenging for a long time due to the difficulty of precisely computing the log-likelihood gradient. Over the past few decades, many works have proposed more or less successful train…
View article: An Introduction to Machine Learning: a perspective from Statistical Physics
An Introduction to Machine Learning: a perspective from Statistical Physics Open
International audience