Differential geometry with extreme eigenvalues in the positive semidefinite cone
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866929376481771520 |
|---|---|
| author | Mostajeran, Cyrus Da Costa, Nathaël Van Goffrier, Graham Sepulchre, Rodolphe |
| author_facet | Mostajeran, Cyrus Da Costa, Nathaël Van Goffrier, Graham Sepulchre, Rodolphe |
| contents | Differential geometric approaches to the analysis and processing of data in the form of symmetric positive definite (SPD) matrices have had notable successful applications to numerous fields including computer vision, medical imaging, and machine learning. The dominant geometric paradigm for such applications has consisted of a few Riemannian geometries associated with spectral computations that are costly at high scale and in high dimensions. We present a route to a scalable geometric framework for the analysis and processing of SPD-valued data based on the efficient computation of extreme generalized eigenvalues through the Hilbert and Thompson geometries of the semidefinite cone. We explore a particular geodesic space structure based on Thompson geometry in detail and establish several properties associated with this structure. Furthermore, we define a novel iterative mean of SPD matrices based on this geometry and prove its existence and uniqueness for a given finite collection of points. Finally, we state and prove a number of desirable properties that are satisfied by this mean. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2304_07347 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Differential geometry with extreme eigenvalues in the positive semidefinite cone Mostajeran, Cyrus Da Costa, Nathaël Van Goffrier, Graham Sepulchre, Rodolphe Differential Geometry Computation Machine Learning Differential geometric approaches to the analysis and processing of data in the form of symmetric positive definite (SPD) matrices have had notable successful applications to numerous fields including computer vision, medical imaging, and machine learning. The dominant geometric paradigm for such applications has consisted of a few Riemannian geometries associated with spectral computations that are costly at high scale and in high dimensions. We present a route to a scalable geometric framework for the analysis and processing of SPD-valued data based on the efficient computation of extreme generalized eigenvalues through the Hilbert and Thompson geometries of the semidefinite cone. We explore a particular geodesic space structure based on Thompson geometry in detail and establish several properties associated with this structure. Furthermore, we define a novel iterative mean of SPD matrices based on this geometry and prove its existence and uniqueness for a given finite collection of points. Finally, we state and prove a number of desirable properties that are satisfied by this mean. |
| title | Differential geometry with extreme eigenvalues in the positive semidefinite cone |
| topic | Differential Geometry Computation Machine Learning |
| url | https://arxiv.org/abs/2304.07347 |