Fast Bayesian experimental design: Laplace-based importance sampling for the expected information gain
For researchers in Bayesian experimental design, this method offers a more efficient computation of expected information gain, addressing the high cost of inner-loop sampling.
The paper presents a computationally efficient method for optimal Bayesian experimental design using Laplace-based importance sampling in the inner loop, reducing computational cost compared to classical double-loop Monte Carlo and a single-loop Laplace approximation method, demonstrated on three numerical examples including sensor placement for electrical impedance tomography.
In calculating expected information gain in optimal Bayesian experimental design, the computation of the inner loop in the classical double-loop Monte Carlo requires a large number of samples and suffers from underflow if the number of samples is small. These drawbacks can be avoided by using an importance sampling approach. We present a computationally efficient method for optimal Bayesian experimental design that introduces importance sampling based on the Laplace method to the inner loop. We derive the optimal values for the method parameters in which the average computational cost is minimized according to the desired error tolerance. We use three numerical examples to demonstrate the computational efficiency of our method compared with the classical double-loop Monte Carlo, and a more recent single-loop Monte Carlo method that uses the Laplace method as an approximation of the return value of the inner loop. The first example is a scalar problem that is linear in the uncertain parameter. The second example is a nonlinear scalar problem. The third example deals with the optimal sensor placement for an electrical impedance tomography experiment to recover the fiber orientation in laminate composites.