Pierre Comon
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Pierre Comon has a total of 70 co-authors. Here are the top co-authors by number of shared publications:
- Konstantin Usevich
Research Center for Automatic Control, French National Center for Scientific Research (CNRS), Universit de Lorraine, Nancy, France
8 shared publications - Jianze Li
Shenzhen Research Institute of Big Data The Chinese University of Hong Kong Shenzhen China
7 shared publications - Vicente Zarzoso
I3S Laboratory, UNS/CNRS, Sophia Antipolis Cedex France 06903
5 shared publications - Elaheh Sobhani
GIPSA-lab, Univ. Grenoble Alpes, CNRS, Grenoble, France
3 shared publications - Khalid Minaoui
LRIT Laboratory, Faculty of Sciences, Mohammed V University, Rabat, Morocco
3 shared publications - Lucas Drumetz
IMT Atlantique, Lab-STICC, UMR 6285, 29238, CNRS, Brest, France
3 shared publications
Latest Publications
- 16 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 36 (9) , 4664-4677
Feature Space Recovery for Efficient Incomplete Multi-View Clustering
T-SVD based incomplete multi-view clustering (IMVC) has received wide attention due to its ability to capture high-order correlations. However, t-SVD suffers from rotation sensitivity, failing to fully explore both inter- and intra-view consistencies. Besides, current methods mainly consider inter- or intra-view correlations, ignoring the low-rank information of sample features within views. To address these weaknesses, we first propose a feature space recovery based IMVC (FSR-IMVC) method, where low-rank feature space recovery and low-rank tensor ring based consistency learning are considered into a unified framework. Furthermore, we extend FSR-IMVC by incorporating anchor learning on the latent feature space, resulting in a scalable FSR-IMVC (sFSR-IMVC) approach that is well-suited to large-scale data. In an iterative way, the learned inter- and intra-view correlations will guide the recovery of missing features, while the explored low-rank information from feature spaces will in turn facilitate consistency exploration, eventually achieving outstanding clustering performance. Experimental results show that FSR-IMVC provides a significant improvement over known state-of-the-art algorithms in terms of ACC, NMI and Purity. Compared with FSR-IMVC, sFSR-IMVC performs slightly worse in clustering accuracy, but offers a notable advantage in computational efficiency, particularly for large-scale datasets. The codes of FSR-IMVC and sFSR-IMVC are publicly available at https://github.com/longzhen520/sFSR-IMVC.
- 16 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 36 (9) , 4664-4677
Feature Space Recovery for Efficient Incomplete Multi-View Clustering
T-SVD based incomplete multi-view clustering (IMVC) has received wide attention due to its ability to capture high-order correlations. However, t-SVD suffers from rotation sensitivity, failing to fully explore both inter- and intra-view consistencies. Besides, current methods mainly consider inter- or intra-view correlations, ignoring the low-rank information of sample features within views. To address these weaknesses, we first propose a feature space recovery based IMVC (FSR-IMVC) method, where low-rank feature space recovery and low-rank tensor ring based consistency learning are considered into a unified framework. Furthermore, we extend FSR-IMVC by incorporating anchor learning on the latent feature space, resulting in a scalable FSR-IMVC (sFSR-IMVC) approach that is well-suited to large-scale data. In an iterative way, the learned inter- and intra-view correlations will guide the recovery of missing features, while the explored low-rank information from feature spaces will in turn facilitate consistency exploration, eventually achieving outstanding clustering performance. Experimental results show that FSR-IMVC provides a significant improvement over known state-of-the-art algorithms in terms of ACC, NMI and Purity. Compared with FSR-IMVC, sFSR-IMVC performs slightly worse in clustering accuracy, but offers a notable advantage in computational efficiency, particularly for large-scale datasets. The codes of FSR-IMVC and sFSR-IMVC are publicly available at https://github.com/longzhen520/sFSR-IMVC.
- Open Access
MultiHU-TD: Multifeature Hyperspectral Unmixing Based on Tensor Decomposition
Hyperspectral unmixing allows representing mixed pixels as a set of pure materials weighted by their abundances. Spectral features alone are often insufficient, so it is common to rely on other features of the scene. Matrix models become insufficient when the hyperspectral image (HSI) is represented as a high-order tensor with additional features in a multimodal, multifeature framework. Tensor models such as canonical polyadic decomposition allow for this kind of unmixing but lack a general framework and interpretability of the results. In this article, we propose an interpretable methodological framework for low-rank multifeature hyperspectral unmixing based on tensor decomposition (MultiHU-TD) that incorporates the abundance sum-to-one constraint in the alternating optimization alternating direction method of multipliers (ADMM) algorithm and provide in-depth mathematical, physical, and graphical interpretation and connections with the extended linear mixing model. As additional features, we propose to incorporate mathematical morphology and reframe a previous work on neighborhood patches within MultiHU-TD. Experiments on real HSIs showcase the interpretability of the model and the analysis of the results. Python and MATLAB implementations are made available on GitHub.
Open AccessMultiHU-TD: Multifeature Hyperspectral Unmixing Based on Tensor Decomposition
Hyperspectral unmixing allows representing mixed pixels as a set of pure materials weighted by their abundances. Spectral features alone are often insufficient, so it is common to rely on other features of the scene. Matrix models become insufficient when the hyperspectral image (HSI) is represented as a high-order tensor with additional features in a multimodal, multifeature framework. Tensor models such as canonical polyadic decomposition allow for this kind of unmixing but lack a general framework and interpretability of the results. In this article, we propose an interpretable methodological framework for low-rank multifeature hyperspectral unmixing based on tensor decomposition (MultiHU-TD) that incorporates the abundance sum-to-one constraint in the alternating optimization alternating direction method of multipliers (ADMM) algorithm and provide in-depth mathematical, physical, and graphical interpretation and connections with the extended linear mixing model. As additional features, we propose to incorporate mathematical morphology and reframe a previous work on neighborhood patches within MultiHU-TD. Experiments on real HSIs showcase the interpretability of the model and the analysis of the results. Python and MATLAB implementations are made available on GitHub.
- 14 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-21
MultiHU-TD: Multifeature Hyperspectral Unmixing Based on Tensor Decomposition
Hyperspectral unmixing allows to represent mixed pixels as a set of pure materials weighted by their abundances. Spectral features alone are often insufficient, so it is common to rely on other features of the scene. Matrix models become insufficient when the hyperspectral image is represented as a high-order tensor with additional features in a multimodal, multi-feature framework. Tensor models such as Canonical polyadic decomposition allow for this kind of unmixing, but lack a general framework and interpretability of the results. In this paper, we propose an interpretable methodological framework for low-rank Multi-feature hyperspectral unmixing based on tensor decomposition (MultiHU-TD) which incorporates the abundance sum-to-one constraint in the Alternating optimization ADMM algorithm, and provide in-depth mathematical, physical and graphical interpretation and connections with the extended linear mixing model. As additional features, we propose to incorporate mathematical morphology and reframe a previous work on neighborhood patches within MultiHU-TD. Experiments on real hyperspectral images showcase the interpretability of the model and the analysis of the results. Python and MATLAB implementations are made available on GitHub.
- 14 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-21
MultiHU-TD: Multifeature Hyperspectral Unmixing Based on Tensor Decomposition
Hyperspectral unmixing allows to represent mixed pixels as a set of pure materials weighted by their abundances. Spectral features alone are often insufficient, so it is common to rely on other features of the scene. Matrix models become insufficient when the hyperspectral image is represented as a high-order tensor with additional features in a multimodal, multi-feature framework. Tensor models such as Canonical polyadic decomposition allow for this kind of unmixing, but lack a general framework and interpretability of the results. In this paper, we propose an interpretable methodological framework for low-rank Multi-feature hyperspectral unmixing based on tensor decomposition (MultiHU-TD) which incorporates the abundance sum-to-one constraint in the Alternating optimization ADMM algorithm, and provide in-depth mathematical, physical and graphical interpretation and connections with the extended linear mixing model. As additional features, we propose to incorporate mathematical morphology and reframe a previous work on neighborhood patches within MultiHU-TD. Experiments on real hyperspectral images showcase the interpretability of the model and the analysis of the results. Python and MATLAB implementations are made available on GitHub.
- 11 May 2023
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Matrix Analysis and Applications
- Vol. 44 (2) , 592-621
Convergence of Gradient-Based Block Coordinate Descent Algorithms for Nonorthogonal Joint Approximate Diagonalization of Matrices
We propose a gradient-based block coordinate descent (BCD-G) framework to solve the joint approximate diagonalization of matrices defined on the product of the complex Stiefel manifold and the special linear group. Instead of the cyclic fashion, we choose a block optimization based on the Riemannian gradient. To update the first block variable in the complex Stiefel manifold, we use the well-known line search descent method. To update the second block variable in the special linear group, based on four different kinds of elementary transformations, we construct three classes, GLU, GQU and GU, and then get three BCD-G algorithms, BCD-GLU, BCD-GQU and BCD-GU. We establish the global convergence and weak convergence of these three algorithms using the Łojasiewicz gradient inequality under the assumption that the iterates are bounded. We also propose a gradient-based Jacobi-type framework to solve the joint approximate diagonalization of matrices defined on the special linear group. As in the BCD-G case, using the GLU and GQU classes of elementary transformations, we focus on the Jacobi-GLU and Jacobi-GQU algorithms and establish their global convergence and weak convergence as well. All the algorithms and convergence results described in this paper also apply to the real case.
- 11 May 2023
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Matrix Analysis and Applications
- Vol. 44 (2) , 592-621
Convergence of Gradient-Based Block Coordinate Descent Algorithms for Nonorthogonal Joint Approximate Diagonalization of Matrices
We propose a gradient-based block coordinate descent (BCD-G) framework to solve the joint approximate diagonalization of matrices defined on the product of the complex Stiefel manifold and the special linear group. Instead of the cyclic fashion, we choose a block optimization based on the Riemannian gradient. To update the first block variable in the complex Stiefel manifold, we use the well-known line search descent method. To update the second block variable in the special linear group, based on four different kinds of elementary transformations, we construct three classes, GLU, GQU and GU, and then get three BCD-G algorithms, BCD-GLU, BCD-GQU and BCD-GU. We establish the global convergence and weak convergence of these three algorithms using the Łojasiewicz gradient inequality under the assumption that the iterates are bounded. We also propose a gradient-based Jacobi-type framework to solve the joint approximate diagonalization of matrices defined on the special linear group. As in the BCD-G case, using the GLU and GQU classes of elementary transformations, we focus on the Jacobi-GLU and Jacobi-GQU algorithms and establish their global convergence and weak convergence as well. All the algorithms and convergence results described in this paper also apply to the real case.
- Open Access
Convergence of gradient-based block coordinate descent algorithms for non-orthogonal joint approximate diagonalization of matrices
In this paper, we propose a gradient-based block coordinate descent (BCD-G) framework to solve the joint approximate diagonalization of matrices defined on the product of the complex Stiefel manifold and the special linear group. Instead of the cyclic fashion, we choose a block optimization based on the Riemannian gradient. To update the first block variable in the complex Stiefel manifold, we use the well-known line search descent method. To update the second block variable in the special linear group, based on four kinds of different elementary transformations, we construct three classes: GLU, GQU and GU, and then get three BCD-G algorithms: BCD-GLU, BCD-GQU and BCD-GU. We establish the global and weak convergence of these three algorithms using the \L{}ojasiewicz gradient inequality under the assumption that the iterates are bounded. We also propose a gradient-based Jacobi-type framework to solve the joint approximate diagonalization of matrices defined on the special linear group. As in the BCD-G case, using the GLU and GQU classes of elementary transformations, we focus on the Jacobi-GLU and Jacobi-GQU algorithms and establish their global and weak convergence. All the algorithms and convergence results described in this paper also apply to the real case.
Open AccessConvergence of gradient-based block coordinate descent algorithms for non-orthogonal joint approximate diagonalization of matrices
In this paper, we propose a gradient-based block coordinate descent (BCD-G) framework to solve the joint approximate diagonalization of matrices defined on the product of the complex Stiefel manifold and the special linear group. Instead of the cyclic fashion, we choose a block optimization based on the Riemannian gradient. To update the first block variable in the complex Stiefel manifold, we use the well-known line search descent method. To update the second block variable in the special linear group, based on four kinds of different elementary transformations, we construct three classes: GLU, GQU and GU, and then get three BCD-G algorithms: BCD-GLU, BCD-GQU and BCD-GU. We establish the global and weak convergence of these three algorithms using the \L{}ojasiewicz gradient inequality under the assumption that the iterates are bounded. We also propose a gradient-based Jacobi-type framework to solve the joint approximate diagonalization of matrices defined on the special linear group. As in the BCD-G case, using the GLU and GQU classes of elementary transformations, we focus on the Jacobi-GLU and Jacobi-GQU algorithms and establish their global and weak convergence. All the algorithms and convergence results described in this paper also apply to the real case.
- Open Access
SFBS: a Forward-Backward Splitting algorithm for constrained tensor decomposition
Tensors (multi-way arrays) and constrained tensor decomposition are very practical in various applications. The existing constrained decomposition algorithms, which are based on Alternating Direction Method of Multipliers (ADMM) or proximal methods, suffer either from a lack of complete convergence guarantee or from the lack of expected accuracy. In this paper, we propose a constrained decomposition algorithm, called SFBS, which stands for ``Simple Forward-Backward Splitting'' and is based on a particular proximal method to handle constraints. SFBS is theoretically and practically ahead compared to the state-of-the-art, since (i) not only SFBS achieves state-of-the-art performances, but also has a complete convergence guarantee, unlike ADMM-based algorithms; (ii) SFBS is much more robust against additive noise and computationally less expensive; (iii) unlike some existing algorithms, SFBS requires to adjust fewer hyperparameters, which are easy to set according to the convergence condition.
Open AccessSFBS: a Forward-Backward Splitting algorithm for constrained tensor decomposition
Tensors (multi-way arrays) and constrained tensor decomposition are very practical in various applications. The existing constrained decomposition algorithms, which are based on Alternating Direction Method of Multipliers (ADMM) or proximal methods, suffer either from a lack of complete convergence guarantee or from the lack of expected accuracy. In this paper, we propose a constrained decomposition algorithm, called SFBS, which stands for ``Simple Forward-Backward Splitting'' and is based on a particular proximal method to handle constraints. SFBS is theoretically and practically ahead compared to the state-of-the-art, since (i) not only SFBS achieves state-of-the-art performances, but also has a complete convergence guarantee, unlike ADMM-based algorithms; (ii) SFBS is much more robust against additive noise and computationally less expensive; (iii) unlike some existing algorithms, SFBS requires to adjust fewer hyperparameters, which are easy to set according to the convergence condition.
- 3 August 2022
- Published by Elsevier in Digital Signal Processing
- Vol. 129
Computation of the nonnegative canonical tensor decomposition with two accelerated proximal gradient algorithms
- 3 August 2022
- Published by Elsevier in Digital Signal Processing
- Vol. 129
Computation of the nonnegative canonical tensor decomposition with two accelerated proximal gradient algorithms
- 27 July 2022
- Published by Elsevier in Signal Processing
- Vol. 201
A normality test for multivariate dependent samples
- 27 July 2022
- Published by Elsevier in Signal Processing
- Vol. 201
A normality test for multivariate dependent samples
- Open Access
A Random Matrix Perspective on Random Tensors
Tensor models play an increasingly prominent role in many fields, notably in machine learning. In several applications, such as community detection, topic modeling and Gaussian mixture learning, one must estimate a low-rank signal from a noisy tensor. Hence, understanding the fundamental limits of estimators of that signal inevitably calls for the study of random tensors. Substantial progress has been recently achieved on this subject in the large-dimensional limit. Yet, some of the most significant among these results--in particular, a precise characterization of the abrupt phase transition (with respect to signal-to-noise ratio) that governs the performance of the maximum likelihood (ML) estimator of a symmetric rank-one model with Gaussian noise--were derived based of mean-field spin glass theory, which is not easily accessible to non-experts. In this work, we develop a sharply distinct and more elementary approach, relying on standard but powerful tools brought by years of advances in random matrix theory. The key idea is to study the spectra of random matrices arising from contractions of a given random tensor. We show how this gives access to spectral properties of the random tensor itself. For the aforementioned rank-one model, our technique yields a hitherto unknown fixed-point equation whose solution precisely matches the asymptotic performance of the ML estimator above the phase transition threshold in the third-order case. A numerical verification provides evidence that the same holds for orders 4 and 5, leading us to conjecture that, for any order, our fixed-point equation is equivalent to the known characterization of the ML estimation performance that had been obtained by relying on spin glasses. Moreover, our approach sheds light on certain properties of the ML problem landscape in large dimensions and can be extended to other models, such as asymmetric and non-Gaussian.
Open AccessA Random Matrix Perspective on Random Tensors
Tensor models play an increasingly prominent role in many fields, notably in machine learning. In several applications, such as community detection, topic modeling and Gaussian mixture learning, one must estimate a low-rank signal from a noisy tensor. Hence, understanding the fundamental limits of estimators of that signal inevitably calls for the study of random tensors. Substantial progress has been recently achieved on this subject in the large-dimensional limit. Yet, some of the most significant among these results--in particular, a precise characterization of the abrupt phase transition (with respect to signal-to-noise ratio) that governs the performance of the maximum likelihood (ML) estimator of a symmetric rank-one model with Gaussian noise--were derived based of mean-field spin glass theory, which is not easily accessible to non-experts. In this work, we develop a sharply distinct and more elementary approach, relying on standard but powerful tools brought by years of advances in random matrix theory. The key idea is to study the spectra of random matrices arising from contractions of a given random tensor. We show how this gives access to spectral properties of the random tensor itself. For the aforementioned rank-one model, our technique yields a hitherto unknown fixed-point equation whose solution precisely matches the asymptotic performance of the ML estimator above the phase transition threshold in the third-order case. A numerical verification provides evidence that the same holds for orders 4 and 5, leading us to conjecture that, for any order, our fixed-point equation is equivalent to the known characterization of the ML estimation performance that had been obtained by relying on spin glasses. Moreover, our approach sheds light on certain properties of the ML problem landscape in large dimensions and can be extended to other models, such as asymmetric and non-Gaussian.
- Open Access
Multivariate Normality test for colored data
Performances of the Multivariate Kurtosis are investigated when applied to colored data, with or without Auto-Regressive pre-whitening, and with or without projection onto a lower-dimensional random subspace. Computer experiments demonstrate the importance of taking into account the possible color of the process in calculating the power of the normality test, in all the scenarios.
Open AccessMultivariate Normality test for colored data
Performances of the Multivariate Kurtosis are investigated when applied to colored data, with or without Auto-Regressive pre-whitening, and with or without projection onto a lower-dimensional random subspace. Computer experiments demonstrate the importance of taking into account the possible color of the process in calculating the power of the normality test, in all the scenarios.