Optimization via Strategic Law of Large Numbers
Xiaohong Chen, Zengjing Chen, Wayne Yuan Gao, Xiaodong Yan, Guodong Zhang · 2024
Read Optimization via Strategic Law of Large Numbers by Xiaohong Chen, Zengjing Chen, Wayne Yuan Gao, Xiaodong Yan, Guodong Zhang (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-2412-05604.txt.
Description
Five researchers, led by Xiaohong Chen and Zengjing Chen, presented a math paper in 2024 on a new way to find the best value of a hard function. Their work links the search task to a two-armed choice model. In that model, a smart sampling plan can push the sample mean toward the true best point.
The team builds a set of methods called Strategic Monte Carlo Optimization. The methods judge each step by the sign of the slope, so no hard equations need to be solved. Tests on tough test functions, and on neural nets with square and hinge loss, reached the global peak with few starts and few steps. The methods beat many well known global search tools in those tests.
Catalog record
| CREATOR | Xiaohong Chen, Zengjing Chen, Wayne Yuan Gao, Xiaodong Yan, Guodong Zhang |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 197 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-05604