Variational Estimators for Bayesian Optimal Experimental Design

03/13/2019 ∙ by Adam Foster, et al. ∙ 16 ∙ share

Bayesian optimal experimental design (BOED) is a principled framework for making efficient use of limited experimental resources. Unfortunately, its applicability is hampered by the difficulty of obtaining accurate estimates of the expected information gain (EIG) of an experiment. To address this, we introduce several classes of fast EIG estimators suited to the experiment design context by building on ideas from variational inference and mutual information estimation. We show theoretically and empirically that these estimators can provide significant gains in speed and accuracy over previous approaches. We demonstrate the practicality of our approach via a number of experiments, including an adaptive experiment with human participants.

READ FULL TEXT

Variational Bayesian Optimal Experimental Design

Bayesian optimal experimental design (BOED) is a principled framework fo...
03/13/2019 ∙ by Adam Foster, et al. ∙ 0 ∙ share

read it
Deep Adaptive Design: Amortizing Sequential Bayesian Experimental Design

We introduce Deep Adaptive Design (DAD), a general method for amortizing...
03/03/2021 ∙ by Adam Foster, et al. ∙ 0 ∙ share

read it
A Unified Stochastic Gradient Approach to Designing Bayesian-Optimal Experiments

We introduce a fully stochastic gradient based approach to Bayesian opti...
11/01/2019 ∙ by Adam Foster, et al. ∙ 23 ∙ share

read it
Sequential Bayesian Experimental Design for Implicit Models via Mutual Information

Bayesian experimental design (BED) is a framework that uses statistical ...
03/20/2020 ∙ by Steven Kleinegesse, et al. ∙ 0 ∙ share

read it
Efficient Bayesian Experimental Design for Implicit Models

Bayesian experimental design involves the optimal allocation of resource...
10/23/2018 ∙ by Steven Kleinegesse, et al. ∙ 0 ∙ share

read it
Neural-Network Heuristics for Adaptive Bayesian Quantum Estimation

Quantum metrology promises unprecedented measurement precision but suffe...
03/04/2020 ∙ by Lukas J. Fiderer, et al. ∙ 0 ∙ share

read it
Bayesian Experimental Design for Implicit Models by Mutual Information Neural Estimation

Implicit stochastic models, where the data-generation distribution is in...
02/19/2020 ∙ by Steven Kleinegesse, et al. ∙ 36 ∙ share

read it

Variational Estimators for Bayesian Optimal Experimental Design

Authors

Variational Bayesian Optimal Experimental Design

Deep Adaptive Design: Amortizing Sequential Bayesian Experimental Design

A Unified Stochastic Gradient Approach to Designing Bayesian-Optimal Experiments

Sequential Bayesian Experimental Design for Implicit Models via Mutual Information

Efficient Bayesian Experimental Design for Implicit Models

Neural-Network Heuristics for Adaptive Bayesian Quantum Estimation

Bayesian Experimental Design for Implicit Models by Mutual Information Neural Estimation

This week in AI