Bayesian Optimization with Preference Exploration using a Monotonic Neural Network Ensemble
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911375514664960 |
|---|---|
| author | Wang, Hanyang Branke, Juergen Poloczek, Matthias |
| author_facet | Wang, Hanyang Branke, Juergen Poloczek, Matthias |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_18792 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Bayesian Optimization with Preference Exploration using a Monotonic Neural Network Ensemble Wang, Hanyang Branke, Juergen Poloczek, Matthias Machine Learning Optimization and Control 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. |
| title | Bayesian Optimization with Preference Exploration using a Monotonic Neural Network Ensemble |
| topic | Machine Learning Optimization and Control |
| url | https://arxiv.org/abs/2501.18792 |