Hybrid Summary Statistics
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908557215006720 |
|---|---|
| author | Makinen, T. Lucas Sui, Ce Wandelt, Benjamin D. Porqueres, Natalia Heavens, Alan |
| author_facet | Makinen, T. Lucas Sui, Ce Wandelt, Benjamin D. Porqueres, Natalia Heavens, Alan |
| contents | We present a way to capture high-information posteriors from training sets that are sparsely sampled over the parameter space for robust simulation-based inference. In physical inference problems, we can often apply domain knowledge to define traditional summary statistics to capture some of the information in a dataset. We show that augmenting these statistics with neural network outputs to maximise the mutual information improves information extraction compared to neural summaries alone or their concatenation to existing summaries and makes inference robust in settings with low training data. We introduce 1) two loss formalisms to achieve this and 2) apply the technique to two different cosmological datasets to extract non-Gaussian parameter information. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_07548 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Hybrid Summary Statistics Makinen, T. Lucas Sui, Ce Wandelt, Benjamin D. Porqueres, Natalia Heavens, Alan Machine Learning Cosmology and Nongalactic Astrophysics Information Theory Data Analysis, Statistics and Probability We present a way to capture high-information posteriors from training sets that are sparsely sampled over the parameter space for robust simulation-based inference. In physical inference problems, we can often apply domain knowledge to define traditional summary statistics to capture some of the information in a dataset. We show that augmenting these statistics with neural network outputs to maximise the mutual information improves information extraction compared to neural summaries alone or their concatenation to existing summaries and makes inference robust in settings with low training data. We introduce 1) two loss formalisms to achieve this and 2) apply the technique to two different cosmological datasets to extract non-Gaussian parameter information. |
| title | Hybrid Summary Statistics |
| topic | Machine Learning Cosmology and Nongalactic Astrophysics Information Theory Data Analysis, Statistics and Probability |
| url | https://arxiv.org/abs/2410.07548 |