PABBO: Preferential Amortized Black-Box Optimization
This addresses the problem of slow PBO computations for users in interactive design optimization, making it more practical for real-world applications, though it is incremental as it builds on amortized BO advances.
The paper tackled the computational inefficiency of Preferential Bayesian Optimization (PBO) by proposing PABBO, a fully amortized method that meta-learns both the surrogate and acquisition functions, resulting in orders of magnitude faster performance and often higher accuracy on synthetic and real-world benchmarks.
Preferential Bayesian Optimization (PBO) is a sample-efficient method to learn latent user utilities from preferential feedback over a pair of designs. It relies on a statistical surrogate model for the latent function, usually a Gaussian process, and an acquisition strategy to select the next candidate pair to get user feedback on. Due to the non-conjugacy of the associated likelihood, every PBO step requires a significant amount of computations with various approximate inference techniques. This computational overhead is incompatible with the way humans interact with computers, hindering the use of PBO in real-world cases. Building on the recent advances of amortized BO, we propose to circumvent this issue by fully amortizing PBO, meta-learning both the surrogate and the acquisition function. Our method comprises a novel transformer neural process architecture, trained using reinforcement learning and tailored auxiliary losses. On a benchmark composed of synthetic and real-world datasets, our method is several orders of magnitude faster than the usual Gaussian process-based strategies and often outperforms them in accuracy.