Re-Envisioning Numerical Information Field Theory (NIFTy.re): A Library for Gaussian Processes and Variational Inference
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917694636294144 |
|---|---|
| author | Edenhofer, Gordian Frank, Philipp Roth, Jakob Leike, Reimar H. Guerdi, Massin Scheel-Platz, Lukas I. Guardiani, Matteo Eberle, Vincent Westerkamp, Margret Enßlin, Torsten A. |
| author_facet | Edenhofer, Gordian Frank, Philipp Roth, Jakob Leike, Reimar H. Guerdi, Massin Scheel-Platz, Lukas I. Guardiani, Matteo Eberle, Vincent Westerkamp, Margret Enßlin, Torsten A. |
| contents | Imaging is the process of transforming noisy, incomplete data into a space that humans can interpret. NIFTy is a Bayesian framework for imaging and has already successfully been applied to many fields in astrophysics. Previous design decisions held the performance and the development of methods in NIFTy back. We present a rewrite of NIFTy, coined NIFTy.re, which reworks the modeling principle, extends the inference strategies, and outsources much of the heavy lifting to JAX. The rewrite dramatically accelerates models written in NIFTy, lays the foundation for new types of inference machineries, improves maintainability, and enables interoperability between NIFTy and the JAX machine learning ecosystem. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_16683 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Re-Envisioning Numerical Information Field Theory (NIFTy.re): A Library for Gaussian Processes and Variational Inference Edenhofer, Gordian Frank, Philipp Roth, Jakob Leike, Reimar H. Guerdi, Massin Scheel-Platz, Lukas I. Guardiani, Matteo Eberle, Vincent Westerkamp, Margret Enßlin, Torsten A. Instrumentation and Methods for Astrophysics Machine Learning Imaging is the process of transforming noisy, incomplete data into a space that humans can interpret. NIFTy is a Bayesian framework for imaging and has already successfully been applied to many fields in astrophysics. Previous design decisions held the performance and the development of methods in NIFTy back. We present a rewrite of NIFTy, coined NIFTy.re, which reworks the modeling principle, extends the inference strategies, and outsources much of the heavy lifting to JAX. The rewrite dramatically accelerates models written in NIFTy, lays the foundation for new types of inference machineries, improves maintainability, and enables interoperability between NIFTy and the JAX machine learning ecosystem. |
| title | Re-Envisioning Numerical Information Field Theory (NIFTy.re): A Library for Gaussian Processes and Variational Inference |
| topic | Instrumentation and Methods for Astrophysics Machine Learning |
| url | https://arxiv.org/abs/2402.16683 |