Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Lenka Zdeborová

Lenka Zdeborová

Stochasticity helps to navigate rough landscapes: comparing gradient-descent-based algorithms in the phase retrieval problem


Mar 08, 2021
Francesca Mignacco, Pierfrancesco Urbani, Lenka Zdeborová

Figure 1 for Stochasticity helps to navigate rough landscapes: comparing gradient-descent-based algorithms in the phase retrieval problem
Figure 2 for Stochasticity helps to navigate rough landscapes: comparing gradient-descent-based algorithms in the phase retrieval problem
Figure 3 for Stochasticity helps to navigate rough landscapes: comparing gradient-descent-based algorithms in the phase retrieval problem
Figure 4 for Stochasticity helps to navigate rough landscapes: comparing gradient-descent-based algorithms in the phase retrieval problem
* 24 pages, 10 figures 
  Access Paper or Ask Questions

Classifying high-dimensional Gaussian mixtures: Where kernel methods fail and neural networks succeed


Feb 23, 2021
Maria Refinetti, Sebastian Goldt, Florent Krzakala, Lenka Zdeborová

Figure 1 for Classifying high-dimensional Gaussian mixtures: Where kernel methods fail and neural networks succeed
Figure 2 for Classifying high-dimensional Gaussian mixtures: Where kernel methods fail and neural networks succeed
Figure 3 for Classifying high-dimensional Gaussian mixtures: Where kernel methods fail and neural networks succeed
Figure 4 for Classifying high-dimensional Gaussian mixtures: Where kernel methods fail and neural networks succeed
* The accompanying code for this paper is available at https://github.com/mariaref/rfvs2lnn_GMM_online 
  Access Paper or Ask Questions

Capturing the learning curves of generic features maps for realistic data sets with a teacher-student model


Feb 16, 2021
Bruno Loureiro, Cédric Gerbelot, Hugo Cui, Sebastian Goldt, Florent Krzakala, Marc Mézard, Lenka Zdeborová

Figure 1 for Capturing the learning curves of generic features maps for realistic data sets with a teacher-student model
Figure 2 for Capturing the learning curves of generic features maps for realistic data sets with a teacher-student model
Figure 3 for Capturing the learning curves of generic features maps for realistic data sets with a teacher-student model
Figure 4 for Capturing the learning curves of generic features maps for realistic data sets with a teacher-student model
* main: 13 pages, 5 figures; appendix: 52 pages, 4 figures 
  Access Paper or Ask Questions

Solvable Model for Inheriting the Regularization through Knowledge Distillation


Dec 02, 2020
Luca Saglietti, Lenka Zdeborová

Figure 1 for Solvable Model for Inheriting the Regularization through Knowledge Distillation
Figure 2 for Solvable Model for Inheriting the Regularization through Knowledge Distillation
Figure 3 for Solvable Model for Inheriting the Regularization through Knowledge Distillation
Figure 4 for Solvable Model for Inheriting the Regularization through Knowledge Distillation

  Access Paper or Ask Questions

Epidemic mitigation by statistical inference from contact tracing data


Sep 20, 2020
Antoine Baker, Indaco Biazzo, Alfredo Braunstein, Giovanni Catania, Luca Dall'Asta, Alessandro Ingrosso, Florent Krzakala, Fabio Mazza, Marc Mézard, Anna Paola Muntoni, Maria Refinetti, Stefano Sarao Mannelli, Lenka Zdeborová

Figure 1 for Epidemic mitigation by statistical inference from contact tracing data
Figure 2 for Epidemic mitigation by statistical inference from contact tracing data
Figure 3 for Epidemic mitigation by statistical inference from contact tracing data
Figure 4 for Epidemic mitigation by statistical inference from contact tracing data
* 21 pages, 7 figures 
  Access Paper or Ask Questions

Optimization and Generalization of Shallow Neural Networks with Quadratic Activation Functions


Jun 27, 2020
Stefano Sarao Mannelli, Eric Vanden-Eijnden, Lenka Zdeborová

Figure 1 for Optimization and Generalization of Shallow Neural Networks with Quadratic Activation Functions
Figure 2 for Optimization and Generalization of Shallow Neural Networks with Quadratic Activation Functions
Figure 3 for Optimization and Generalization of Shallow Neural Networks with Quadratic Activation Functions
Figure 4 for Optimization and Generalization of Shallow Neural Networks with Quadratic Activation Functions
* 10 pages, 4 figures + appendix 
  Access Paper or Ask Questions

The Gaussian equivalence of generative models for learning with two-layer neural networks


Jun 25, 2020
Sebastian Goldt, Galen Reeves, Marc Mézard, Florent Krzakala, Lenka Zdeborová

Figure 1 for The Gaussian equivalence of generative models for learning with two-layer neural networks
Figure 2 for The Gaussian equivalence of generative models for learning with two-layer neural networks
Figure 3 for The Gaussian equivalence of generative models for learning with two-layer neural networks
Figure 4 for The Gaussian equivalence of generative models for learning with two-layer neural networks
* The accompanying code for this paper is available at https://github.com/sgoldt/gaussian-equiv-2layer 
  Access Paper or Ask Questions

Complex Dynamics in Simple Neural Networks: Understanding Gradient Flow in Phase Retrieval


Jun 12, 2020
Stefano Sarao Mannelli, Giulio Biroli, Chiara Cammarota, Florent Krzakala, Pierfrancesco Urbani, Lenka Zdeborová

Figure 1 for Complex Dynamics in Simple Neural Networks: Understanding Gradient Flow in Phase Retrieval
Figure 2 for Complex Dynamics in Simple Neural Networks: Understanding Gradient Flow in Phase Retrieval
Figure 3 for Complex Dynamics in Simple Neural Networks: Understanding Gradient Flow in Phase Retrieval
Figure 4 for Complex Dynamics in Simple Neural Networks: Understanding Gradient Flow in Phase Retrieval
* 9 pages, 5 figures + appendix 
  Access Paper or Ask Questions

Generalization error in high-dimensional perceptrons: Approaching Bayes error with convex optimization


Jun 11, 2020
Benjamin Aubin, Florent Krzakala, Yue M. Lu, Lenka Zdeborová

Figure 1 for Generalization error in high-dimensional perceptrons: Approaching Bayes error with convex optimization
Figure 2 for Generalization error in high-dimensional perceptrons: Approaching Bayes error with convex optimization
Figure 3 for Generalization error in high-dimensional perceptrons: Approaching Bayes error with convex optimization
Figure 4 for Generalization error in high-dimensional perceptrons: Approaching Bayes error with convex optimization
* 11 pages + 45 pages Supplementary Material / 5 figures 
  Access Paper or Ask Questions

Dynamical mean-field theory for stochastic gradient descent in Gaussian mixture classification


Jun 10, 2020
Francesca Mignacco, Florent Krzakala, Pierfrancesco Urbani, Lenka Zdeborová

Figure 1 for Dynamical mean-field theory for stochastic gradient descent in Gaussian mixture classification
Figure 2 for Dynamical mean-field theory for stochastic gradient descent in Gaussian mixture classification
Figure 3 for Dynamical mean-field theory for stochastic gradient descent in Gaussian mixture classification
Figure 4 for Dynamical mean-field theory for stochastic gradient descent in Gaussian mixture classification
* 8 pages + appendix, 4 figures 
  Access Paper or Ask Questions

Phase retrieval in high dimensions: Statistical and computational phase transitions


Jun 09, 2020
Antoine Maillard, Bruno Loureiro, Florent Krzakala, Lenka Zdeborová

Figure 1 for Phase retrieval in high dimensions: Statistical and computational phase transitions
Figure 2 for Phase retrieval in high dimensions: Statistical and computational phase transitions
Figure 3 for Phase retrieval in high dimensions: Statistical and computational phase transitions
Figure 4 for Phase retrieval in high dimensions: Statistical and computational phase transitions
* 11 pages (main text and references), 26 pages of supplementary material 
  Access Paper or Ask Questions

TRAMP: Compositional Inference with TRee Approximate Message Passing


Apr 03, 2020
Antoine Baker, Benjamin Aubin, Florent Krzakala, Lenka Zdeborová

Figure 1 for TRAMP: Compositional Inference with TRee Approximate Message Passing
Figure 2 for TRAMP: Compositional Inference with TRee Approximate Message Passing
Figure 3 for TRAMP: Compositional Inference with TRee Approximate Message Passing
Figure 4 for TRAMP: Compositional Inference with TRee Approximate Message Passing
* Source code available at https://github.com/sphinxteam/tramp. For some examples, see https://github.com/benjaminaubin/tramp_examples 
  Access Paper or Ask Questions

The role of regularization in classification of high-dimensional noisy Gaussian mixture


Feb 26, 2020
Francesca Mignacco, Florent Krzakala, Yue M. Lu, Lenka Zdeborová

Figure 1 for The role of regularization in classification of high-dimensional noisy Gaussian mixture
Figure 2 for The role of regularization in classification of high-dimensional noisy Gaussian mixture
Figure 3 for The role of regularization in classification of high-dimensional noisy Gaussian mixture
Figure 4 for The role of regularization in classification of high-dimensional noisy Gaussian mixture
* 8 pages + appendix, 6 figures 
  Access Paper or Ask Questions

Generalisation error in learning with random features and the hidden manifold model


Feb 21, 2020
Federica Gerace, Bruno Loureiro, Florent Krzakala, Marc Mézard, Lenka Zdeborová

Figure 1 for Generalisation error in learning with random features and the hidden manifold model
Figure 2 for Generalisation error in learning with random features and the hidden manifold model
Figure 3 for Generalisation error in learning with random features and the hidden manifold model
Figure 4 for Generalisation error in learning with random features and the hidden manifold model

  Access Paper or Ask Questions

Large deviations for the perceptron model and consequences for active learning


Dec 09, 2019
Hugo Cui, Luca Saglietti, Lenka Zdeborová

Figure 1 for Large deviations for the perceptron model and consequences for active learning
Figure 2 for Large deviations for the perceptron model and consequences for active learning
Figure 3 for Large deviations for the perceptron model and consequences for active learning
Figure 4 for Large deviations for the perceptron model and consequences for active learning
* 25 pages, 7 figures 
  Access Paper or Ask Questions

Rademacher complexity and spin glasses: A link between the replica and statistical theories of learning


Dec 05, 2019
Alia Abbara, Benjamin Aubin, Florent Krzakala, Lenka Zdeborová

Figure 1 for Rademacher complexity and spin glasses: A link between the replica and statistical theories of learning
Figure 2 for Rademacher complexity and spin glasses: A link between the replica and statistical theories of learning
Figure 3 for Rademacher complexity and spin glasses: A link between the replica and statistical theories of learning
* 15 + 10 pages 
  Access Paper or Ask Questions

Exact asymptotics for phase retrieval and compressed sensing with random generative priors


Dec 04, 2019
Benjamin Aubin, Bruno Loureiro, Antoine Baker, Florent Krzakala, Lenka Zdeborová

Figure 1 for Exact asymptotics for phase retrieval and compressed sensing with random generative priors
Figure 2 for Exact asymptotics for phase retrieval and compressed sensing with random generative priors
Figure 3 for Exact asymptotics for phase retrieval and compressed sensing with random generative priors
Figure 4 for Exact asymptotics for phase retrieval and compressed sensing with random generative priors
* 12+2 pages, 6 figures 
  Access Paper or Ask Questions

Modelling the influence of data structure on learning in neural networks


Sep 25, 2019
Sebastian Goldt, Marc Mézard, Florent Krzakala, Lenka Zdeborová

Figure 1 for Modelling the influence of data structure on learning in neural networks
Figure 2 for Modelling the influence of data structure on learning in neural networks
Figure 3 for Modelling the influence of data structure on learning in neural networks
Figure 4 for Modelling the influence of data structure on learning in neural networks

  Access Paper or Ask Questions

Who is Afraid of Big Bad Minima? Analysis of Gradient-Flow in a Spiked Matrix-Tensor Model


Jul 22, 2019
Stefano Sarao Mannelli, Giulio Biroli, Chiara Cammarota, Florent Krzakala, Lenka Zdeborová

Figure 1 for Who is Afraid of Big Bad Minima? Analysis of Gradient-Flow in a Spiked Matrix-Tensor Model
Figure 2 for Who is Afraid of Big Bad Minima? Analysis of Gradient-Flow in a Spiked Matrix-Tensor Model
Figure 3 for Who is Afraid of Big Bad Minima? Analysis of Gradient-Flow in a Spiked Matrix-Tensor Model
Figure 4 for Who is Afraid of Big Bad Minima? Analysis of Gradient-Flow in a Spiked Matrix-Tensor Model
* 9 pages, 4 figures + appendix 
  Access Paper or Ask Questions

Dynamics of stochastic gradient descent for two-layer neural networks in the teacher-student setup


Jun 18, 2019
Sebastian Goldt, Madhu S. Advani, Andrew M. Saxe, Florent Krzakala, Lenka Zdeborová

Figure 1 for Dynamics of stochastic gradient descent for two-layer neural networks in the teacher-student setup
Figure 2 for Dynamics of stochastic gradient descent for two-layer neural networks in the teacher-student setup
Figure 3 for Dynamics of stochastic gradient descent for two-layer neural networks in the teacher-student setup
Figure 4 for Dynamics of stochastic gradient descent for two-layer neural networks in the teacher-student setup
* 10 pages + references + supplemental material. arXiv admin note: substantial text overlap with arXiv:1901.09085 
  Access Paper or Ask Questions

On the Universality of Noiseless Linear Estimation with Respect to the Measurement Matrix


Jun 11, 2019
Alia Abbara, Antoine Baker, Florent Krzakala, Lenka Zdeborová

Figure 1 for On the Universality of Noiseless Linear Estimation with Respect to the Measurement Matrix
Figure 2 for On the Universality of Noiseless Linear Estimation with Respect to the Measurement Matrix
Figure 3 for On the Universality of Noiseless Linear Estimation with Respect to the Measurement Matrix
Figure 4 for On the Universality of Noiseless Linear Estimation with Respect to the Measurement Matrix
* 13 pages, 4 figures 
  Access Paper or Ask Questions

The spiked matrix model with generative priors


May 30, 2019
Benjamin Aubin, Bruno Loureiro, Antoine Maillard, Florent Krzakala, Lenka Zdeborová

Figure 1 for The spiked matrix model with generative priors
Figure 2 for The spiked matrix model with generative priors
Figure 3 for The spiked matrix model with generative priors
Figure 4 for The spiked matrix model with generative priors
* 12 + 56, 8 figures, v2 lighter jpeg figures 
  Access Paper or Ask Questions

Passed & Spurious: analysing descent algorithms and local minima in spiked matrix-tensor model


Feb 01, 2019
Stefano Sarao Mannelli, Florent Krzakala, Pierfrancesco Urbani, Lenka Zdeborová

Figure 1 for Passed & Spurious: analysing descent algorithms and local minima in spiked matrix-tensor model
Figure 2 for Passed & Spurious: analysing descent algorithms and local minima in spiked matrix-tensor model
Figure 3 for Passed & Spurious: analysing descent algorithms and local minima in spiked matrix-tensor model
Figure 4 for Passed & Spurious: analysing descent algorithms and local minima in spiked matrix-tensor model
* 12 pages + appendix, 10 figures 
  Access Paper or Ask Questions

Generalisation dynamics of online learning in over-parameterised neural networks


Jan 25, 2019
Sebastian Goldt, Madhu S. Advani, Andrew M. Saxe, Florent Krzakala, Lenka Zdeborová

Figure 1 for Generalisation dynamics of online learning in over-parameterised neural networks
Figure 2 for Generalisation dynamics of online learning in over-parameterised neural networks
Figure 3 for Generalisation dynamics of online learning in over-parameterised neural networks
Figure 4 for Generalisation dynamics of online learning in over-parameterised neural networks
* 25 pages, 13 figures 
  Access Paper or Ask Questions

Marvels and Pitfalls of the Langevin Algorithm in Noisy High-dimensional Inference


Dec 21, 2018
Stefano Sarao Mannelli, Giulio Biroli, Chiara Cammarota, Florent Krzakala, Pierfrancesco Urbani, Lenka Zdeborová

Figure 1 for Marvels and Pitfalls of the Langevin Algorithm in Noisy High-dimensional Inference
Figure 2 for Marvels and Pitfalls of the Langevin Algorithm in Noisy High-dimensional Inference
Figure 3 for Marvels and Pitfalls of the Langevin Algorithm in Noisy High-dimensional Inference
Figure 4 for Marvels and Pitfalls of the Langevin Algorithm in Noisy High-dimensional Inference
* 10 pages and 5 figures + appendix 
  Access Paper or Ask Questions

Rank-one matrix estimation: analysis of algorithmic and information theoretic limits by the spatial coupling method


Dec 06, 2018
Jean Barbier, Mohamad Dia, Nicolas Macris, Florent Krzakala, Lenka Zdeborová

Figure 1 for Rank-one matrix estimation: analysis of algorithmic and information theoretic limits by the spatial coupling method
Figure 2 for Rank-one matrix estimation: analysis of algorithmic and information theoretic limits by the spatial coupling method
Figure 3 for Rank-one matrix estimation: analysis of algorithmic and information theoretic limits by the spatial coupling method
Figure 4 for Rank-one matrix estimation: analysis of algorithmic and information theoretic limits by the spatial coupling method
* Submitted to Journal of Machine Learning Research (JMLR) 
  Access Paper or Ask Questions

Optimal Errors and Phase Transitions in High-Dimensional Generalized Linear Models


Nov 01, 2018
Jean Barbier, Florent Krzakala, Nicolas Macris, Léo Miolane, Lenka Zdeborová

Figure 1 for Optimal Errors and Phase Transitions in High-Dimensional Generalized Linear Models
Figure 2 for Optimal Errors and Phase Transitions in High-Dimensional Generalized Linear Models
* 101 pages, 5 figures 
  Access Paper or Ask Questions

Entropy and mutual information in models of deep neural networks


Oct 29, 2018
Marylou Gabrié, Andre Manoel, Clément Luneau, Jean Barbier, Nicolas Macris, Florent Krzakala, Lenka Zdeborová

Figure 1 for Entropy and mutual information in models of deep neural networks
Figure 2 for Entropy and mutual information in models of deep neural networks
Figure 3 for Entropy and mutual information in models of deep neural networks
Figure 4 for Entropy and mutual information in models of deep neural networks

  Access Paper or Ask Questions

Approximate message-passing for convex optimization with non-separable penalties


Sep 17, 2018
Andre Manoel, Florent Krzakala, Gaël Varoquaux, Bertrand Thirion, Lenka Zdeborová

Figure 1 for Approximate message-passing for convex optimization with non-separable penalties
Figure 2 for Approximate message-passing for convex optimization with non-separable penalties
Figure 3 for Approximate message-passing for convex optimization with non-separable penalties
Figure 4 for Approximate message-passing for convex optimization with non-separable penalties
* 18 pages, 6 figures 
  Access Paper or Ask Questions

Approximate Survey Propagation for Statistical Inference


Jul 03, 2018
Fabrizio Antenucci, Florent Krzakala, Pierfrancesco Urbani, Lenka Zdeborová

Figure 1 for Approximate Survey Propagation for Statistical Inference
Figure 2 for Approximate Survey Propagation for Statistical Inference
Figure 3 for Approximate Survey Propagation for Statistical Inference
Figure 4 for Approximate Survey Propagation for Statistical Inference
* 37 pages, 14 figures 
  Access Paper or Ask Questions