Diversity By Design: Leveraging Distribution Matching for Offline Model-Based Optimization
Michael S. Yao, James C. Gee, Osbert Bastani · 2025
Read Diversity By Design: Leveraging Distribution Matching for Offline Model-Based Optimization by Michael S. Yao, James C. Gee, Osbert Bastani (2025) 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-2501-18768.txt.
Introduction
Michael Yao, James Gee, and Osbert Bastani work in 2025 on offline model-based optimization, where new designs must be proposed using only a fixed dataset of past examples. A good method should not just chase the single best design; it should offer a varied slate of strong candidates.
Their method, DynAMO, treats variety as a matching problem: the spread of proposed designs should mirror the spread already present in the offline data. Framed this way, diversity becomes an explicit goal that can be added to any optimization setup. Tests across several tasks support the idea. Diverse candidates give engineers real choices when the top pick fails.
Catalog record
| CREATOR | Michael S. Yao, James C. Gee, Osbert Bastani |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 211 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-18768