Learn-As-You-Go Acceleration of Cosmological Parameter Estimates
arXiv:1506.01079 · doi:10.1088/1475-7516/2015/09/005
Abstract
Cosmological analyses can be accelerated by approximating slow calculations using a training set, which is either precomputed or generated dynamically. However, this approach is only safe if the approximations are well understood and controlled. This paper surveys issues associated with the use of machine-learning based emulation strategies for accelerating cosmological parameter estimation. We describe a learn-as-you-go algorithm that is implemented in the Cosmo++ code and (1) trains the emulator while simultaneously estimating posterior probabilities; (2) identifies unreliable estimates, computing the exact numerical likelihoods if necessary; and (3) progressively learns and updates the error model as the calculation progresses. We explicitly describe and model the emulation error and show how this can be propagated into the posterior probabilities. We apply these techniques to the Planck likelihood and the calculation of CDM posterior probabilities. The computation is significantly accelerated without a pre-defined training set and uncertainties in the posterior probabilities are subdominant to statistical fluctuations. We have obtained a speedup factor of for Metropolis-Hastings and for nested sampling. Finally, we discuss the general requirements for a credible error model and show how to update them on-the-fly.
arXiv comment20 pages excluding Appendices, 5 figures. Code publicly available at http://cosmopp.com; matches published version
References in corpus (22)
- Nine-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Cosmological Parameter Results
- Efficient Computation of CMB anisotropies in closed FRW models
- Cosmological parameters from CMB and other data: a Monte-Carlo approach
- MultiNest: an efficient and robust Bayesian inference tool for cosmology and particle physics
- The Cosmic Linear Anisotropy Solving System (CLASS) II: Approximation schemes
- Multimodal nested sampling: an efficient and robust alternative to MCMC methods for astronomical data analysis
- Planck 2013 results. XV. CMB power spectra and likelihood
- The Coyote Universe I: Precision Determination of the Nonlinear Matter Power Spectrum
- The Coyote Universe II: Cosmological Models and Precision Emulation of the Nonlinear Matter Power Spectrum
- The Coyote Universe Extended: Precision Emulation of the Matter Power Spectrum
- The Coyote Universe III: Simulation Suite and Precision Emulator for the Nonlinear Matter Power Spectrum
- SKYNET: an efficient and robust neural network training tool for machine learning in astronomy
- Fast cosmological parameter estimation using neural networks
- Cosmic Emulation: The Concentration-Mass Relation for wCDM Universes
- Rapid Calculation of Theoretical CMB Angular Power Spectra
- Rico: An Accurate Cosmological Recombination Code
- Fast Cosmological Parameter Estimation from Microwave Background Temperature and Polarization Power Spectra
- Fast Bayesian inference for slow-roll inflation
- Intelligent Design: On the Emulation of Cosmological Simulations
- Cosmo++: An Object-Oriented C++ Library for Cosmology
- Caching and Interpolated Likelihoods: Accelerating Cosmological Monte Carlo Markov Chains
- Pico: Parameters for the Impatient Cosmologist