Bayesian Optimization with Preference Exploration using a Monotonic Neural Network Ensemble
This work addresses interactive preference learning in multi-objective black-box optimization for real-world applications, but it is incremental as it builds on existing BOPE methods by incorporating monotonicity.
The paper tackled the Bayesian Optimization with Preference Exploration (BOPE) problem by proposing a neural network ensemble as a utility surrogate model that integrates monotonicity and handles pairwise comparison data, resulting in outperforming state-of-the-art approaches and showing robustness to noise in utility evaluations.
Many real-world black-box optimization problems have multiple conflicting objectives. Rather than attempting to approximate the entire set of Pareto-optimal solutions, interactive preference learning allows to focus the search on the most relevant subset. However, few previous studies have exploited the fact that utility functions are usually monotonic. In this paper, we address the Bayesian Optimization with Preference Exploration (BOPE) problem and propose using a neural network ensemble as a utility surrogate model. This approach naturally integrates monotonicity and supports pairwise comparison data. Our experiments demonstrate that the proposed method outperforms state-of-the-art approaches and exhibits robustness to noise in utility evaluations. An ablation study highlights the critical role of monotonicity in enhancing performance.