Show simple item record

contributor authorKamrah, Eesh
contributor authorGhoreishi, Seyede Fatemeh
contributor authorDing, Zijian “Jason”
contributor authorChan, Joel
contributor authorFuge, Mark
date accessioned2023-11-29T19:29:17Z
date available2023-11-29T19:29:17Z
date copyright8/29/2023 12:00:00 AM
date issued8/29/2023 12:00:00 AM
date issued2023-08-29
identifier issn1050-0472
identifier othermd_145_11_111703.pdf
identifier urihttp://yetl.yabesh.ir/yetl1/handle/yetl/4294798
description abstractDesign researchers have struggled to produce quantitative predictions for exactly why and when diversity might help or hinder design search efforts. This paper addresses that problem by studying one ubiquitously used search strategy—Bayesian optimization (BO)—on a 2D test problem with modifiable convexity and difficulty. Specifically, we test how providing diverse versus non-diverse initial samples to BO affects its performance during search and introduce a fast ranked-determinantal point process method for computing diverse sets, which we need to detect sets of highly diverse or non-diverse initial samples. We initially found, to our surprise, that diversity did not appear to affect BO, neither helping nor hurting the optimizer’s convergence. However, follow-on experiments illuminated a key trade-off. Non-diverse initial samples hastened posterior convergence for the underlying model hyper-parameters—a model building advantage. In contrast, diverse initial samples accelerated exploring the function itself—a space exploration advantage. Both advantages help BO, but in different ways, and the initial sample diversity directly modulates how BO trades those advantages. Indeed, we show that fixing the BO hyper-parameters removes the model building advantage, causing diverse initial samples to always outperform models trained with non-diverse samples. These findings shed light on why, at least for BO-type optimizers, the use of diversity has mixed effects and cautions against the ubiquitous use of space-filling initializations in BO. To the extent that humans use explore-exploit search strategies similar to BO, our results provide a testable conjecture for why and when diversity may affect human-subject or design team experiments.
publisherThe American Society of Mechanical Engineers (ASME)
titleHow Diverse Initial Samples Help and Hurt Bayesian Optimizers
typeJournal Paper
journal volume145
journal issue11
journal titleJournal of Mechanical Design
identifier doi10.1115/1.4063006
journal fristpage111703-1
journal lastpage111703-11
page11
treeJournal of Mechanical Design:;2023:;volume( 145 ):;issue: 011
contenttypeFulltext


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record