Hybrid Summary Statistics

Document detail

ID

oai:arXiv.org:2410.07548

Topic

Statistics - Machine Learning Astrophysics - Cosmology and Nonga... Computer Science - Information The... Computer Science - Machine Learnin... Physics - Data Analysis, Statistic...

Author

Makinen, T. Lucas Sui, Ce Wandelt, Benjamin D. Porqueres, Natalia Heavens, Alan

Year

2024

listing date

10/16/2024

Keywords

inference information

Metrics

Abstract

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.

;Comment: 7 pages, 4 figures.

Accepted to ML4PS2024 at NeurIPS 2024

Makinen, T. Lucas,Sui, Ce,Wandelt, Benjamin D.,Porqueres, Natalia,Heavens, Alan, 2024, Hybrid Summary Statistics