Pith. sign in

REVIEW 2 cited by

Fast cosmological parameter estimation using neural networks

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv astro-ph/0608174 v2 pith:U4SFMNRA submitted 2006-08-08 astro-ph

Fast cosmological parameter estimation using neural networks

classification astro-ph
keywords cosmonetpowerspectraflatlikelihoodmodelsparametertimes
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

We present a method for accelerating the calculation of 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 \Lambda CDM models containing the 3\sigma WMAP1 confidence region. We also use CosmoNet to compute the WMAP3 likelihood for flat \Lambda CDM models and show that marginalised 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% of cosmic variance, and that CosmoNet is \sim 7 \times 10^4 faster than CAMB (for flat models) and \sim 6 \times 10^6 times faster than the official WMAP3 likelihood code. CosmoNet and an interface to CosmoMC are publically available at www.mrao.cam.ac.uk/software/cosmonet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. CLiENT: A new tool for emulating cosmological likelihoods using deep neural networks

    astro-ph.CO 2025-12 conditional novelty 6.0

    A neural-network tool that directly emulates cosmological likelihood functions, achieving ~0.1σ-accurate credible intervals with roughly 20,000 likelihood evaluations.

  2. Machine Learning for Multi-messenger Probes of New Physics and Cosmology: A Review and Perspective

    hep-ph 2026-04 unverdicted novelty 3.0

    A review summarizing machine learning methods for multi-messenger probes of dark matter and new physics, with a proposed plan for future integrated analyses.