Abstract:Scientific modeling applications often require estimating a distribution of parameters consistent with a dataset of observations - an inference task also known as source distribution estimation. This problem can be ill-posed, however, since many different source distributions might produce the same distribution of data-consistent simulations. To make a principled choice among many equally valid sources, we propose an approach which targets the maximum entropy distribution, i.e., prioritizes retaining as much uncertainty as possible. Our method is purely sample-based - leveraging the Sliced-Wasserstein distance to measure the discrepancy between the dataset and simulations - and thus suitable for simulators with intractable likelihoods. We benchmark our method on several tasks, and show that it can recover source distributions with substantially higher entropy than recent source estimation methods, without sacrificing the fidelity of the simulations. Finally, to demonstrate the utility of our approach, we infer source distributions for parameters of the Hodgkin-Huxley model from experimental datasets with thousands of single-neuron measurements. In summary, we propose a principled method for inferring source distributions of scientific simulator parameters while retaining as much uncertainty as possible.

Source Function Determined from HBT Correlations by the Maximum Entropy Principle

Source Function Determined from Hanbury-Brown/Twiss Correlations by the Maximum Entropy Principle

Imaging Source with Gaussian Proper Time Distribution

Dual Giant Gravitons in Sasaki-Einstein Backgrounds

An inversion-asymmetric source function for HBT analysis in relativistic heavy ion collisions

Source Function from Two-Particle Correlation Through Deblurring

Measuring the Size of the Source for the Quarks and Gluons Jet with the HBT Correlation Method

An Inverse Random Source Problem for Maxwell's Equations

The Inversion-Asymmetry of Particle Emission Source in Relativistic Heavy-Ion Collisions

An Inverse Random Source Problem for the Helmholtz Equation.

The use of non-local constraints in maximum-entropy electron density reconstruction

Investigating the excitation function of HBT radii for Lévy-stable sources

Sourcerer: Sample-based Maximum Entropy Source Distribution Estimation

Coulomb and strong interactions in the final state of HBT correlations for Lévy type source functions

Uniqueness of an inverse source problem in experimental aeroacoustics

Quantum sources and a quantum coding theorem

Anisotropic Transverse Flow and the HBT Correlation Function

Correspondence between HBT radii and the emission zone in non-central heavy ion collisions

High Frequency Analysis of Helmholtz Equations: Case of Two Point Sources

Combining Bayesian reconstruction entropy with maximum entropy method for analytic continuations of matrix-valued Green's functions

Holographic Derivation of a Class of Short Range Correlation Functions