IA2 · RESEARCH
Research catalogue
AI and advanced inference across astrophysics. Papers from 2024 onward, including preprints, with an IA2 member among the first three listed authors.
66 papers · Checked 2026-09-24 · Updated every two weeks
66 matching papers
Semiresolved Stellar Populations as Distance Indicators
Neural posterior estimation turns partially resolved stellar populations into distance measurements.
Beyond traditional emission-line diagnostics: using autoencoders to uncover active galactic nuclei in DESI spectra
An autoencoder searches DESI spectra for active galactic nuclei beyond conventional emission-line diagnostics.
First detection of ultra-fast outflows in a quiescent galaxy
Bayesian spectral modelling with neural-assisted nested sampling tests evidence for an ultra-fast outflow in a quiescent galaxy.
A Pixel-by-Pixel Path to Population III Discovery with JWST
Pixel-level inference searches JWST images for signatures of Population III stars.
3DSTokesFlow: simulation-based inference for 3D Stokes profiles using flow matching
Flow matching estimates atmospheric parameters from three-dimensional Stokes profiles.
Marginal multi-object multi-frame blind deconvolution
Marginalizing over nuisance parameters improves multi-object, multi-frame restoration of astronomical images.
Accelerating 3D Non-LTE Synthesis with Graph Neural Networks
Graph neural networks accelerate three-dimensional non-LTE spectral synthesis.
IRIS NUV diagnostics for Ellerman bombs: spectral properties, thermodynamics, and formation height
IRIS²⁺ spectral inversions recover atmospheric stratification and formation heights for Ellerman bombs.
A comparison of pendulum models for large-amplitude longitudinal prominence oscillations
Bayesian parameter inference and model comparison quantify prominence magnetic fields under competing pendulum models.
Inspectorch: Efficient rare event exploration in solar observations
Flow-based density estimation identifies rare solar events through probabilistic anomaly scores.
Signatures of Damping Nonlinear Oscillations by KHI-induced Turbulence in Synthetic Observations
Bayesian fits to synthetic loop oscillations constrain amplitudes and periods while exposing degeneracies in damping parameters.
Talking with the Latents -- how to convert your LLM into an astronomer
A latent representation connects astronomical data to a language model for interpreting galaxy properties.
Inferring physical parameters of solar filaments from simultaneous longitudinal and transverse oscillations
Bayesian seismology combines longitudinal and transverse filament oscillations to infer magnetic fields, flux-tube lengths and twists.
Cored galaxies in cuspy dark matter halos
Likelihood-based fits and model comparison test whether cuspy dark matter halos can reproduce apparently cored galaxies.
ELG×LRG distribution through dark matter halo dynamics
A two-level Bayesian framework connects emission-line and luminous red galaxies to dark matter halo dynamics.
Silicate emission in a type-2 quasar: JWST/MIRI constraints on torus geometry and radiative feedback
Bayesian torus modelling uses JWST silicate emission to constrain the geometry of an obscured quasar.
Thermodynamic and magnetic evolution of an eruptive C-class solar flare observed with SST/TRIPPEL-SP
Non-LTE spectropolarimetric inversions reconstruct the atmospheric evolution of an eruptive C-class solar flare.
Differentiable Fuzzy Cosmic-Web for Field Level Inference
A differentiable cosmic-web description supports gradient-based inference of large-scale structure.
Stellar-like Galactic center excess challenges particle dark matter
Adaptive template fitting and MultiNest posterior sampling constrain dark matter contributions to the Galactic center excess.
The spatially-resolved effect of mergers on the stellar mass assembly of MaNGA galaxies
Diffusion-based inference maps how mergers contribute to the spatial distribution of stellar mass in MaNGA galaxies.
Probing the Parameter Space of Axion-Like Particles Using Simulation-Based Inference
Truncated marginal neural ratio estimation constrains axion-like particle parameters from simulated CTAO spectra.
Identification of Nonlinear Damping of Transverse Loop Oscillations by KHI-induced Turbulence
MCMC fitting and Bayesian model comparison test nonlinear turbulence damping against linear models of coronal-loop oscillations.
Neural translation for Stokes inversion and synthesis
A neural translation model connects Stokes spectra and atmospheric structure for both inversion and synthesis.
Beyond Traditional Diagnostics: Identifying Active Galactic Nuclei with Spectral Energy Distribution Fitting in DESI Data
CIGALE spectral energy distribution modelling identifies AGN in DESI data and compares them with traditional selections.
Simulation-based inference of galaxy properties from JWST pixels
Neural inference estimates resolved galaxy properties directly from JWST pixels.
Chromospheric velocities in an M3.2 flare using He I 1083.0 nm and Ca II 854.2 nm
Inversions of helium and calcium spectra recover chromospheric velocities during a solar flare.
torchmfbd: a flexible multi-object multi-frame blind deconvolution code
A flexible image-restoration framework estimates objects and atmospheric aberrations through maximum-a-posteriori optimization.
Automatic detection of Ellerman bombs using Deep Learning
Neural networks detect Ellerman bombs in solar observations and assess the spectral and spatial information needed.
A robust neural determination of the source-count distribution of the Fermi-LAT sky at high latitudes
Neural ratio estimation detects gamma-ray sources and reconstructs source-count distributions from Fermi-LAT data.
Euclid Quick Data Release (Q1) Exploring galaxy properties with a multi-modal foundation model
A multimodal foundation model extracts physical information from Euclid galaxy observations.
Euclid Quick Data Release (Q1), A first look at the fraction of bars in massive galaxies at z<1
Deep-learning morphology measurements characterize the fraction of barred galaxies in Euclid Q1.
Euclid Quick Data Release (Q1): First visual morphology catalogue
A deep-learning pipeline produces the first visual morphology catalogue for Euclid Q1.
Galaxy mass profiles with convolutional neural networks
Convolutional networks recover galaxy mass profiles and quantify uncertainty from simulated observations.
MaNGA AGN dwarf galaxies (MAD). III. The role of mergers and environment in active galactic nucleus activity in dwarf galaxies
CIGALE fits correct stellar masses and pair mass ratios when testing mergers as triggers of dwarf-galaxy AGN.
Study of an active region prominence using spectropolarimetric data in the He I D3 multiplet
Helium-line spectropolarimetric inversions probe the magnetic structure of an active-region prominence.
Cosmology with One Galaxy: Auto-Encoding the Galaxy Properties Manifold
An autoencoder explores the galaxy-property manifold and the cosmological information encoded in individual galaxies.
COSMOS-Web: The emergence of the Hubble Sequence
Deep-learning morphology measurements trace the emergence of the Hubble sequence in COSMOS-Web.
ZF-UDS-7329: A relic galaxy in the early Universe
Full-spectrum stellar-population fitting and Monte Carlo uncertainties reconstruct the formation history of a high-redshift relic galaxy.
A statistical study of lopsided galaxies using random forest
Random forests identify which galaxy properties are associated with lopsided structure.
Spectral resolution effects on the information content in solar spectra
SIR inversions of degraded synthetic Stokes spectra measure how spectral resolution affects inferred solar atmospheric parameters.
Exploring spectropolarimetric inversions using neural fields. Solar chromospheric magnetic field under the weak-field approximation
Neural fields regularize spectropolarimetric inversions of the solar chromospheric magnetic field.
New constraints on the central mass contents of Omega Centauri from combined stellar kinematics and pulsar timing
Nested sampling combines stellar motions and pulsar timing to constrain the central mass components of Omega Centauri.
Bayesian Deconvolution of Astronomical Images with Diffusion Models: Quantifying Prior-Driven Features in Reconstructions
Diffusion priors support Bayesian image deconvolution and reveal which reconstructed features are driven by the prior.
Non-Local Thermodynamic Equilibrium inversions of the Si I 10827 A spectral line
Non-LTE inversions recover solar atmospheric information from the silicon 10827 Å line.
Simulation-based inference of the 2D ex-situ stellar mass fraction distribution of galaxies using variational autoencoders
Variational autoencoders infer two-dimensional maps of stellar mass acquired through galaxy mergers.
No evidence for gamma-ray emission from the Sagittarius dwarf spheroidal galaxy
Adaptive gamma-ray templates and nested sampling of pixel-count statistics test emission associated with the Sagittarius dwarf galaxy.
Value Added Catalog of physical properties of more than 1.3 million galaxies from the DESI Survey
Spectral energy distribution fitting estimates physical properties for more than 1.3 million DESI galaxies.
Properties of sunspot light bridges on a geometric height scale
Spectral inversions reconstruct the atmospheric properties of sunspot light bridges on a geometric height scale.
The negative BAO shift in the Lyα forest from cosmological simulations
Markov-chain Monte Carlo fits quantify the baryon acoustic oscillation shift in simulated Lyman-alpha forests.
Constraints on the in-situ and ex-situ stellar masses in nearby galaxies with Artificial Intelligence
AI estimates how much stellar mass nearby galaxies formed internally and acquired through mergers.
Deriving the star formation histories of galaxies from spectra with simulation-based inference
Simulation-based inference reconstructs galaxy star-formation histories from spectra with posterior uncertainties.
A fast neural emulator for interstellar chemistry
A neural emulator accelerates interstellar chemical calculations for physical modelling and inference.
The PAU Survey: galaxy stellar population properties estimates with narrowband data
CIGALE and Prospector infer galaxy properties from PAU narrow-band photometry and assess model-dependent uncertainties.
Simulation-based inference of radio millisecond pulsars in globular clusters
Marginal neural ratio estimation constrains the size and luminosity distribution of millisecond-pulsar populations in globular clusters.
Fast simulation mapping: from standard to modified gravity cosmologies using the bias assignment method
A calibrated bias-assignment method maps standard cosmological simulations into modified-gravity tracer distributions.
Solar multi-object multi-frame blind deconvolution with a spatially variant convolution neural emulator
A neural emulator accelerates solar image restoration with spatially varying atmospheric degradation.
From VIPERS to SDSS: Unveiling galaxy spectra evolution over 9 Gyr through unsupervised machine-learning
Unsupervised spectral classification compares galaxy populations across nine billion years of evolution.
Probing star formation rates and histories in AGN and non-AGN galaxies across diverse cosmic environments and extensive X-ray luminosity ranges
Likelihood-weighted CIGALE estimates compare star formation in AGN hosts and inactive galaxies across environments.
Robust inference of the Galactic centre gamma-ray excess spatial properties
Bayesian evidence compares Galactic center excess morphologies after adaptive optimization of diffuse-emission backgrounds.
A post-merger enhancement only in star-forming Type 2 Seyfert galaxies: the deep learning view
Deep-learning merger identification tests the connection between galaxy interactions and Seyfert activity.
Global Coronal Magnetic Field Estimation Using Bayesian Inference
Bayesian inference maps coronal magnetic-field strengths from CoMP wave and density measurements, including uncertainty.
Galaxies Going Bananas: Inferring the 3D Geometry of High-Redshift Galaxies with JWST-CEERS
Bayesian geometric inference reconstructs the three-dimensional shapes of high-redshift galaxies from JWST observations.
Bayesian deep learning for cosmic volumes with modified gravity
Bayesian neural networks distinguish modified-gravity cosmologies from cosmic density volumes while quantifying uncertainty.
Uncovering axion-like particles in supernova gamma-ray spectra
MultiNest posterior sampling reconstructs axion-like particle parameters from simulated supernova gamma-ray spectra.
Galaxy Morphology from z≈6 through the eyes of JWST
Deep-learning morphology classification follows galaxy structure to redshift six with JWST.
On the nature of disks at high redshift seen by JWST/CEERS with contrastive learning and cosmological simulations
Contrastive learning compares high-redshift JWST galaxies with cosmological simulations to investigate disk structure.
How papers are selected
At least one current group member must appear in positions 1–3 of the source author list. Collaboration names count as listed entries. We include explicit applications of machine learning, Bayesian or simulation-based inference, and advanced inverse methods; ordinary regression or a future inference application alone does not qualify.
Dates refer to the journal year where available, otherwise the first preprint year. Earlier preprints published since 2024 are included when publication metadata is verified. Preprint and journal versions are combined. The initial collection was checked against abstracts and, where needed, full texts; new entries are selected automatically from explicit method evidence in titles and abstracts.
Coverage depends on source indexing and author-name variants. Methods described only in full text may need a manual addition. Search record and maintenance details ↗