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

Abstract

The cosmic microwave background power spectra are a primary window into the early universe. However, achieving interpretable compression and fast inference diagnostics 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 retaining an interface to likelihood-style diagnostic tests. The model achieves high directional reconstruction fidelity for the $D_{\ell }^{TT}$, $D_{\ell }^{EE}$, and $D_{\ell }^{TE}$ spectra in just 5 latent dimensions. It reconstructs spectra for several beyond-$\Lambda$CDM test cases, including controlled parameter extrapolations, and enables an amortized surrogate diagnostic that reduces one representative post-training MCMC run from $\sim$40 hours on CPU cores to $\sim$2 minutes on a GPU in this demonstration. The learned latent space shows a distributed, partially structured organization that mirrors known cosmological parameters and their degeneracies. It also provides representation-space discrimination diagnostics for distinguishing tested cosmological spectra from a fiducial reference. Overall, this physics-informed CVAE supports interpretable compression, rapid diagnostic exploration, and anomaly-sensitive representation learning beyond $\Lambda$CDM.

22 pages, 18 figures; accepted for publication in Physical Review D

References in corpus (124)

• emcee: The MCMC Hammer
• Detection of the Baryon Acoustic Peak in the Large-Scale Correlation Function of SDSS Luminous Red Galaxies
• f(R) Theories Of Gravity
• Cosmological parameters from CMB and other data: a Monte-Carlo approach
• Exploring the Expansion History of the Universe
• A Comprehensive Measurement of the Local Value of the Hubble Constant with 1 km/s/Mpc Uncertainty from the Hubble Space Telescope and the SH0ES Team
• In the Realm of the Hubble tension $-$ a Review of Solutions
• Models of f(R) Cosmic Acceleration that Evade Solar-System Tests
• f(T) teleparallel gravity and cosmology
• Bayes in the sky: Bayesian inference and model selection in cosmology
• Planck 2018 results. V. CMB power spectra and likelihoods
• Early Dark Energy Can Resolve The Hubble Tension
• The trouble with $H_0$
• Cobaya: Code for Bayesian Analysis of hierarchical physical models
• Permutation p-values should never be zero: calculating exact p-values when permutations are randomly drawn
• DESI DR2 Results II: Measurements of Baryon Acoustic Oscillations and Cosmological Constraints
• CMB power spectrum parameter degeneracies in the era of precision cosmology
• New physics in light of the $H_0$ tension: an alternative view
• Planck 2018 results. VII. Isotropy and Statistics of the CMB
• Interacting Dark Matter and Dark Energy
• Holographic Dark Energy
• Constraints on holographic dark energy from type Ia supernova observations
• Can interacting dark energy solve the $H_0$ tension?
• Seven hints that early-time new physics alone is not sufficient to solve the Hubble tension
• Interacting dark energy in the early 2020s: a promising solution to the $H_0$ and cosmic shear tensions
• CMB Anomalies after Planck
• Cosmology with Interaction between Phantom Dark Energy and Dark Matter and the Coincidence Problem
• Statefinder diagnostic for holographic dark energy model
• The CosmoVerse White Paper: Addressing observational tensions in cosmology with systematics and fundamental physics
• Supernova Constraints on a holographic dark energy model
• Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks
• Real-time gravitational-wave science with neural posterior estimation
• Matching matched filtering with deep networks in gravitational-wave astronomy
• Statefinder diagnostic for coupled quintessence
• Constraints on Holographic Dark Energy from Latest Supernovae, Galaxy Clustering, and Cosmic Microwave Background Anisotropy Observations
• Interacting scenarios with dynamical dark energy: observational constraints and alleviation of the $H_0$ tension
• Dynamical vacuum energy, holographic quintom, and the reconstruction of scalar-field dark energy
• Bayesian parameter estimation using conditional variational autoencoders for gravitational-wave astronomy
• Interacting Dark Energy and Dark Matter: observational Constraints from Cosmological Parameters
• Consistency tests of $\Lambda$CDM from the early integrated Sachs-Wolfe effect: Implications for early-time new physics and the Hubble tension
• Holographic Ricci dark energy: Current observational constraints, quintom feature, and the reconstruction of scalar-field dark energy
• New generalized Chaplygin gas as a scheme for unification of dark energy and dark matter
• COSMOPOWER: emulating cosmological power spectra for accelerated Bayesian inference from next-generation surveys
• Can the $H_0$ tension be resolved in extensions to $\Lambda$CDM cosmology?
• Testing the Interaction between Dark Energy and Dark Matter with Planck Data
• Kaniadakis holographic dark energy and cosmology
• Coupled Quintessence in a Power-Law Case and the Cosmic Coincidence Problem
• Constraints on the dark energy from holography
• Learning Bayesian posteriors with neural networks for gravitational-wave inference
• Constraints on Interacting Dark Energy Models from the DESI Baryon Acoustic Oscillation and DES Supernovae Data
• A measurement of Hubble's Constant using Fast Radio Bursts
• Implications for the Hubble tension from the ages of the oldest astrophysical objects
• Neural Importance Sampling for Rapid and Reliable Gravitational-Wave Inference
• Field theoretic interpretations of interacting dark energy scenarios and recent observations
• Large-scale stable interacting dark energy model: Cosmological perturbations and observational constraints
• Parametrized Post-Friedmann Framework for Interacting Dark Energy
• Chameleon dark energy can resolve the Hubble tension
• Interacting Dark Energy in a closed universe
• Reduced-order modeling with artificial neurons for gravitational-wave inference
• Exploring the full parameter space for an interacting dark energy model with recent observations including redshift-space distortions: Application of the parametrized post-Friedmann approach
• Holographic dark energy from minimal supergravity
• MGCAMB with massive neutrinos and dynamical dark energy
• Accelerated, Scalable and Reproducible AI-driven Gravitational Wave Detection
• Revisit of the interacting holographic dark energy model after Planck 2015
• Impacts of dark energy on weighing neutrinos after Planck 2015
• Applying deep neural networks to the detection and space parameter estimation of compact binary coalescence with a network of gravitational wave detectors
• Prospects for improving cosmological parameter estimation with gravitational-wave standard sirens from Taiji
• Revisit of the Interaction between Holographic Dark Energy and Dark Matter
• Holographic dark energy in a universe with spatial curvature and massive neutrinos: a full Markov Chain Monte Carlo exploration
• Cosmological constraints on neutrinos after BICEP2
• Neutrinos in the holographic dark energy model: constraints from latest measurements of expansion history and growth of structure
• The Taiji-TianQin-LISA network: Precisely measuring the Hubble constant using both bright and dark sirens
• Synergy between CSST galaxy survey and gravitational-wave observation: Inferring the Hubble constant from dark standard sirens
• Real-time gravitational-wave inference for binary neutron stars using machine learning
• Probing the sign-changeable interaction between dark energy and dark matter with DESI baryon acoustic oscillations and DES supernovae data
• Full-sky Cosmic Microwave Background Foreground Cleaning Using Machine Learning
• Tightening the reins on non-minimal dark sector physics: Interacting Dark Energy with dynamical and non-dynamical equation of state
• Prospects for measuring the Hubble constant and dark energy using gravitational-wave dark sirens with neutron star tidal deformation
• Normalizing flows as an avenue to study overlapping gravitational wave signals
• Prospects for probing the interaction between dark energy and dark matter using gravitational-wave dark sirens with neutron star tidal deformation
• Cosmology with fast radio bursts in the era of SKA
• CosmoPower-JAX: high-dimensional Bayesian inference with differentiable cosmological emulators
• A new coupled three-form dark energy model and implications for the $H_0$ tension
• Space-based gravitational wave signal detection and extraction with deep neural network
• Impacts of dark energy on weighing neutrinos after DESI BAO
• ECoPANN: A Framework for Estimating Cosmological Parameters using Artificial Neural Networks
• GWSkyNet: a real-time classifier for public gravitational-wave candidates
• Recovering the CMB Signal with Machine Learning
• Dispu$\tau$able: the high cost of a low optical depth
• A comprehensive forecast for cosmological parameter estimation using joint observations of gravitational waves and short $\gamma$-ray bursts
• The 21-cm forest as a simultaneous probe of dark matter and cosmic heating history
• Efficient parameter inference for gravitational wave signals in the presence of transient noises using temporal and time-spectral fusion normalizing flow
• Advancing Space-Based Gravitational Wave Astronomy: Rapid Parameter Estimation via Normalizing Flows
• Fast and realistic large-scale structure from machine-learning-augmented random field simulations
• Machine learning cosmic backreaction and its effects on observations
• Interacting dark energy after DESI DR2: a challenge for $\Lambda$CDM paradigm?
• The future of cosmological likelihood-based inference: accelerated high-dimensional parameter estimation and model comparison
• Exploring the Key Features of Repeating Fast Radio Bursts with Machine Learning
• Explaining dark matter halo density profiles with neural networks
• Model-independent $H_0$ within FLRW: Joint constraints from GWTC-3 standard sirens and strong lensing time delays
• syren-new: Precise formulae for the linear and nonlinear matter power spectra with massive neutrinos and dynamical dark energy
• From Images to Features: Unbiased Morphology Classification via Variational Auto-Encoders and Domain Adaptation
• Recovering Cosmic Microwave Background Polarization Signals with Machine Learning
• Narrowing down the Hubble tension to the first two rungs of distance ladders
• Parameter estimation of microlensed gravitational waves with Conditional Variational Autoencoders
• A representation learning approach to probe for dynamical dark energy in matter power spectra
• Delensing of Cosmic Microwave Background Polarization with machine learning
• Distinguishing Coupled Dark Energy Models with Neural Networks
• Deep learning-driven likelihood-free parameter inference for 21-cm forest observations
• cp3-bench: A tool for benchmarking symbolic regression algorithms tested with cosmology
• Combining strongly lensed and unlensed fast radio bursts: To be a more precise late-universe probe
• Dark Energy in the DESI Era: A Brief Review of Evidence, Beyond-$\Lambda$CDM Interpretations, and Tensions
• Deep learning insights into non-universality in the halo mass function
• Psi-GAN: A power-spectrum-informed generative adversarial network for the emulation of large-scale structure maps across cosmologies and redshifts
• Enhancing Cosmological Model Selection with Interpretable Machine Learning
• Cosmological parameter constraints using phenomenological symbolic expressions: On the significance of symbolic expression complexity and accuracy
• Accelerating Bayesian Sampling for Massive Black Hole Binaries with Prior Constraints from Conditional Variational Autoencoder
• $\Lambda$CDM and early dark energy in latent space: a data-driven parametrization of the CMB temperature power spectrum
• Cosmology with One Galaxy: Auto-Encoding the Galaxy Properties Manifold
• Cosmic Kite: Auto-encoding the Cosmic Microwave Background
• CMB delensing with deep learning
• One latent to fit them all: a unified representation of baryonic feedback on matter distribution
• Forecast constraints on null tests of the $\Lambda$CDM model with SPHEREx
• Neural Networks for cosmological model selection and feature importance using Cosmic Microwave Background data