Simulation-based inference with scattering representations: scattering is all you need
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909411333636096 |
|---|---|
| author | Lin, Kiyam Joachimi, Benjamin McEwen, Jason D. |
| author_facet | Lin, Kiyam Joachimi, Benjamin McEwen, Jason D. |
| contents | We demonstrate the successful use of scattering representations without further compression for simulation-based inference (SBI) with images (i.e. field-level), illustrated with a cosmological case study. Scattering representations provide a highly effective representational space for subsequent learning tasks, although the higher dimensional compressed space introduces challenges. We overcome these through spatial averaging, coupled with more expressive density estimators. Compared to alternative methods, such an approach does not require additional simulations for either training or computing derivatives, is interpretable, and resilient to covariate shift. As expected, we show that a scattering only approach extracts more information than traditional second order summary statistics. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_11883 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Simulation-based inference with scattering representations: scattering is all you need Lin, Kiyam Joachimi, Benjamin McEwen, Jason D. Machine Learning Cosmology and Nongalactic Astrophysics Instrumentation and Methods for Astrophysics We demonstrate the successful use of scattering representations without further compression for simulation-based inference (SBI) with images (i.e. field-level), illustrated with a cosmological case study. Scattering representations provide a highly effective representational space for subsequent learning tasks, although the higher dimensional compressed space introduces challenges. We overcome these through spatial averaging, coupled with more expressive density estimators. Compared to alternative methods, such an approach does not require additional simulations for either training or computing derivatives, is interpretable, and resilient to covariate shift. As expected, we show that a scattering only approach extracts more information than traditional second order summary statistics. |
| title | Simulation-based inference with scattering representations: scattering is all you need |
| topic | Machine Learning Cosmology and Nongalactic Astrophysics Instrumentation and Methods for Astrophysics |
| url | https://arxiv.org/abs/2410.11883 |