alpha-Deep Probabilistic Inference (alpha-DPI): efficient uncertainty quantification from exoplanet astrometry to black hole feature extraction

Abstract

Inference is crucial in modern astronomical research, where hidden astrophysical features and patterns are often estimated from indirect and noisy measurements. Inferring the posterior of hidden features, conditioned on the observed measurements, is essential for understanding the uncertainty of results and downstream scientific interpretations. Traditional approaches for posterior estimation include sampling-based methods and variational inference. However, sampling-based methods are typically slow for high-dimensional inverse problems, while variational inference often lacks estimation accuracy. In this paper, we propose alpha-DPI, a deep learning framework that first learns an approximate posterior using alpha-divergence variational inference paired with a generative neural network, and then produces more accurate posterior samples through importance re-weighting of the network samples. It inherits strengths from both sampling and variational inference methods: it is fast, accurate, and scalable to high-dimensional problems. We apply our approach to two high-impact astronomical inference problems using real data: exoplanet astrometry and black hole feature extraction.

References in corpus (22)

• Astropy: A Community Python Package for Astronomy
• Binary companions of evolved stars in APOGEE DR14: Search method and catalog of ~5,000 companions
• dynesty: A Dynamic Nested Sampling Package for Estimating Bayesian Posteriors and Evidences
• First M87 Event Horizon Telescope Results. VI. The Shadow and Mass of the Central Black Hole
• Rényi Divergence and Kullback-Leibler Divergence
• First M87 Event Horizon Telescope Results. IV. Imaging the Central Supermassive Black Hole
• First M87 Event Horizon Telescope Results. II. Array and Instrumentation
• Chaotic star formation and the alignment of stellar rotation with disc and planetary orbital axes
• Dynamic temperature selection for parallel-tempering in Markov chain Monte Carlo simulations
• Interferometric Imaging Directly with Closure Phases and Closure Amplitudes
• orbitize!: A Comprehensive Orbit-fitting Software Package for the High-contrast Imaging Community
• Orbits for the Impatient: A Bayesian Rejection Sampling Method for Quickly Fitting the Orbits of Long-Period Exoplanets
• Evaluation of New Submillimeter VLBI Sites for the Event Horizon Telescope
• The Orbit and Transit Prospects for $\beta$ Pictoris b constrained with One Milliarcsecond Astrometry
• The mass of Beta Pictoris c from Beta Pictoris b orbital motion
• Resonant capture by inward migrating planets
• Long period planets from dynamical relaxation
• Rényi Divergence Variational Inference
• Superresolution Interferometric Imaging with Sparse Modeling Using Total Squared Variation --- Application to Imaging the Black Hole Shadow
• Hybrid Very Long Baseline Interferometry Imaging and Modeling with Themis
• Hint of curvature in the orbital motion of the exoplanet 51 Eridani b using 3 years of VLT/SPHERE monitoring
• Biases in orbit fitting of directly-imaged exoplanets with small orbital coverage

Cited by in corpus (6)

• Key Science Goals for the Next-Generation Event Horizon Telescope
• Neural Importance Sampling for Rapid and Reliable Gravitational-Wave Inference
• Event-horizon-scale Imaging of M87* under Different Assumptions via Deep Generative Image Priors
• Deep learning inference with the Event Horizon Telescope II. The Zingularity framework for Bayesian artificial neural networks
• Interferometric Image Reconstruction using Closure Invariants and Machine Learning
• Accelerating the CLEAN algorithm of radio interferometry with convex optimization