Non-Myopic Multi-Objective Bayesian Optimization
Syrine Belakaria* [email protected], Alaleh Ahmadianshalchi* [email protected], Barbara E Engelhardt [email protected], Stefano Ermon [email protected], Janardhan Rao Doppa [email protected] · 2024
Read Non-Myopic Multi-Objective Bayesian Optimization by Syrine Belakaria* [email protected], Alaleh Ahmadianshalchi* [email protected], Barbara E Engelhardt [email protected], Stefano Ermon [email protected], Janardhan Rao Doppa [email protected] (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-08085.txt.
Description
Syrine Belakaria and four co-authors introduced non-myopic methods in twenty twenty-four for multi-objective Bayesian optimization, where each experiment is expensive and several goals compete. Think materials discovery on a small lab budget: choosing tests one at a time, greedily, wastes the horizon. Single-objective lookahead methods lean on Bellman optimality, which breaks for most multi-objective rewards.
Their workaround uses hypervolume improvement as a scalar reward, letting a lower bound on the Bellman equation stand in for true lookahead. From it they derive three acquisition functions: an exact nested one, a cheaper joint lower bound, and a fast batch-based approximation. On diverse real-world problems the non-myopic functions substantially beat standard myopic ones, showing foresight pays even when objectives conflict.
Catalog record
| CREATOR | Syrine Belakaria* [email protected], Alaleh Ahmadianshalchi* [email protected], Barbara E Engelhardt [email protected], Stefano Ermon [email protected], Janardhan Rao Doppa [email protected] |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-08085