skip to content

Faculty of Economics

Journal Cover

Auld, T., Bridges, M., Hobson, M. P. and Gull, S. F.

Fast cosmological parameter estimation using neural networks

Notices of the Royal Astronomical Society

Vol. 376(1) pp. L11-L15 (2007)

Abstract: We present a method for accelerating the calculation of cosmic microwave background (CMB) power spectra, matter power spectra and likelihood functions for use in cosmological parameter estimation. The algorithm, called COSMONET, is based on training a multilayer perceptron neural network and shares all the advantages of the recently released PICO algorithm of Fendt & Wandelt, but has several additional benefits in terms of simplicity, computational speed, memory requirements and ease of training. We demonstrate the capabilities of COSMONET by computing CMB power spectra over a box in the parameter space of flat [code] cold dark matter (ΛCDM) models containing the [code] WMAP 1-year confidence region. We also use COSMONET to compute the WMAP 3-year (WMAP3) likelihood for flat ΛCDM models and show that marginalized posteriors on parameters derived are very similar to those obtained using CAMB and the WMAP3 code. We find that the average error in the power spectra is typically 2–3 per cent of cosmic variance, and that COSMONET is [code] faster than CAMB (for flat models) and [code] times faster than the official WMAP3 likelihood code.

Keywords: methods: data analysis, methods: statistical, cosmology: cosmic microwave background

Author links: Thomas Auld  

Publisher's Link:

Open Data link:

Papers and Publications

Recent Publications

Jochmans, K. and Weidner, M. Fixed-Effect Regressions on Network Data Econometrica [2019]

Onatski, A. and Wang, C. Spurious Factor Analysis Econometrica, forthcoming [2020]

Elliott, M., Georg, C-P. and Hazell, J. Systemic Risk Shifting in Financial Networks Journal of Economic Theory [2021]

Vogt, M., Linton, O. Multiscale clustering of nonparametric regression curves Journal of Econometrics [2020]