The future of cosmological likelihood-based inference: accelerated high-dimensional parameter estimation and model comparison
Problem
Bayesian parameter estimation and model comparison in cosmology becomes computationally slow in high-dimensional settings, limiting analyses.
Why it matters
Slow inference constrains how quickly cosmologists can extract results from large observational datasets.
What the AI did
A machine learning emulator (CosmoPower-JAX) mimics cosmological observables, combined with differentiable/probabilistic programming (JAX, NumPyro), gradient-based MCMC, and learned harmonic mean evidence estimation to accelerate full Bayesian analysis.
Results
No concrete numbers; abstract claims analysis in 'a fraction of the time' of traditional approaches
Impact
Demonstrated a complete accelerated Bayesian analysis pipeline, but on simulated data per the abstract.
Record
| Venue | The Open Journal of Astrophysics |
|---|---|
| Published | 2024 |
| Work type | article |
| Authors | Davide Piras, Alicja Polanska, A. Spurio Mancini, Matthew A. Price, Jason D. McEwen |
| Citations | 19 (OpenAlex, at import time) |
| Model or method | CosmoPower-JAX neural network emulator; JAX/NumPyro; harmonic |
| Organization | not stated in source |
| Primary source | https://doi.org/10.33232/001c.123368 |
| Open access | diamond |
| Keywords | cosmology, Bayesian inference, machine learning emulation, MCMC, model selection |
Importance rating
real but narrower advance
Accelerates Bayesian cosmological inference, cutting a simulated Stage-IV analysis from 8 months on 48 CPUs to 2 days on 12 GPUs; benefits cosmology analyses, demonstrated on simulated data.
More in Physics & Astronomy
| # | Paper | Year | Citations | Importance |
|---|---|---|---|---|
| 1 | Magnetic control of tokamak plasmas through deep reinforcement learning | 2022 | 739 |
7
|
| 2 | Solving the quantum many-body problem with artificial neural networks | 2017 | 2,205 |
7
|
| 3 | Learning high-accuracy error decoding for quantum processors | 2024 | 90 |
7
|
| 4 | Optimizing Quantum Error Correction Codes with Reinforcement Learning | 2019 | 158 |
7
|
| 5 | Machine-learning-assisted correction of correlated qubit errors in a topological code | 2018 | 141 |
7
|