Multiobjective Tree-Structured Parzen Estimator

Practitioners often encounter challenging real-world problems that involve a simultaneous optimization of multiple objectives in complex search space. To address these problems, we propose practical multiobjective Bayesian algorithm. It is an extension the widely used Tree-structured Parzen Estimator (TPE) algorithm, called Multiobjective (MOTPE). We demonstrate MOTPE approximates Pareto fronts variety benchmark and convolutional neural network design problem better than existing methods through numerical results. also investigate how configuration affects behavior performance method effectiveness asynchronous parallelization based on empirical