Lazier ABC
This work addresses computational efficiency for researchers using approximate Bayesian computation, but it is incremental as it builds on prior lazy ABC methods.
The paper tackles the high computational cost of ABC algorithms by summarizing and extending the lazy ABC algorithm, which reduces computational demand by abandoning unpromising simulations early while maintaining the target distribution through reweighting, and extends it to non-uniform ABC kernels to simplify tuning.
ABC algorithms involve a large number of simulations from the model of interest, which can be very computationally costly. This paper summarises the lazy ABC algorithm of Prangle (2015), which reduces the computational demand by abandoning many unpromising simulations before completion. By using a random stopping decision and reweighting the output sample appropriately, the target distribution is the same as for standard ABC. Lazy ABC is also extended here to the case of non-uniform ABC kernels, which is shown to simplify the process of tuning the algorithm effectively.