English

Conditional variational autoencoders for cosmological model discrimination and anomaly detection in cosmic microwave background power spectra

Cosmology and Nongalactic Astrophysics 2025-11-03 v1 Instrumentation and Methods for Astrophysics General Relativity and Quantum Cosmology High Energy Physics - Phenomenology

Abstract

The cosmic microwave background power spectra are a primary window into the early universe. However, achieving interpretable, likelihood-compatible compression and fast inference under weak model assumptions remains challenging. We propose a parameter-conditioned variational autoencoder (CVAE) that aligns a data-driven latent representation with cosmological parameters while remaining compatible with standard likelihood analyses. The model achieves high-fidelity compression of the DTTD_\ell^{TT}, DEED_\ell^{EE}, and DTED_\ell^{TE} spectra into just 5 latent dimensions, with reconstruction accuracy exceeding 99.9%99.9\% within Planck uncertainties. It reliably reconstructs spectra for beyond-Λ\LambdaCDM scenarios, even under parameter extrapolation, and enables rapid inference, reducing the computation time from \sim40 hours to \sim2 minutes while maintaining posterior consistency. The learned latent space demonstrates a physically meaningful structure, capturing a distributed representation that mirrors known cosmological parameters and their degeneracies. Moreover, it supports highly effective unsupervised discrimination among cosmological models, achieving performance competitive with supervised approaches. Overall, this physics-informed CVAE enables anomaly detection beyond Λ\LambdaCDM and points to physically meaningful directions for refinement.

Keywords

Cite

@article{arxiv.2510.27086,
  title  = {Conditional variational autoencoders for cosmological model discrimination and anomaly detection in cosmic microwave background power spectra},
  author = {Tian-Yang Sun and Tian-Nuo Li and He Wang and Jing-Fei Zhang and Xin Zhang},
  journal= {arXiv preprint arXiv:2510.27086},
  year   = {2025}
}

Comments

19 pages, 13 figures