Bayesian Sequential Optimal Experimental Design
A decision problem in which measurements are selected sequentially within a Bayesian framework to maximize expected information gain. Because the design space suffers from combinatorial explosion, conventional optimization methods are computationally infeasible.
Timeline (last 90 days)
A recent work reformulates the Bayesian sequential optimal experimental design (sOED) problem as a reinforcement learning (RL) task for adaptive measurement in single-pixel imaging.
Traditional methods directly optimize the Bayesian sOED objective function to obtain the next design, which is the main bottleneck faced by this approach and is computationally almost infeasible.
Traditional sOED requires re-solving an optimization problem for the current posterior at each measurement, with computational cost accumulating over measurement steps; the RL method instead shifts the computational burden to the offline training stage, with only policy inference at the online stage.
All Facts (3)
A recent work reformulates the Bayesian sequential optimal experimental design (sOED) problem as a reinforcement learning (RL) task for adaptive measurement in single-pixel imaging.
50%UnverifiedTraditional methods directly optimize the Bayesian sOED objective function to obtain the next design, which is the main bottleneck faced by this approach and is computationally almost infeasible.
50%UnverifiedTraditional sOED requires re-solving an optimization problem for the current posterior at each measurement, with computational cost accumulating over measurement steps; the RL method instead shifts the computational burden to the offline training stage, with only policy inference at the online stage.
50%