Local Search GFlowNets

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Kim, Minsu, Yun, Taeyoung, Bengio, Emmanuel, Zhang, Dinghuai, Bengio, Yoshua, Ahn, Sungsoo, Park, Jinkyoo
Format:	Preprint
Published:	2023
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866917620546011136
author	Kim, Minsu Yun, Taeyoung Bengio, Emmanuel Zhang, Dinghuai Bengio, Yoshua Ahn, Sungsoo Park, Jinkyoo
author_facet	Kim, Minsu Yun, Taeyoung Bengio, Emmanuel Zhang, Dinghuai Bengio, Yoshua Ahn, Sungsoo Park, Jinkyoo
contents	Generative Flow Networks (GFlowNets) are amortized sampling methods that learn a distribution over discrete objects proportional to their rewards. GFlowNets exhibit a remarkable ability to generate diverse samples, yet occasionally struggle to consistently produce samples with high rewards due to over-exploration on wide sample space. This paper proposes to train GFlowNets with local search, which focuses on exploiting high-rewarded sample space to resolve this issue. Our main idea is to explore the local neighborhood via backtracking and reconstruction guided by backward and forward policies, respectively. This allows biasing the samples toward high-reward solutions, which is not possible for a typical GFlowNet solution generation scheme, which uses the forward policy to generate the solution from scratch. Extensive experiments demonstrate a remarkable performance improvement in several biochemical tasks. Source code is available: \url{https://github.com/dbsxodud-11/ls_gfn}.
format	Preprint
id	arxiv_https___arxiv_org_abs_2310_02710
institution	arXiv
publishDate	2023
record_format	arxiv
spellingShingle	Local Search GFlowNets Kim, Minsu Yun, Taeyoung Bengio, Emmanuel Zhang, Dinghuai Bengio, Yoshua Ahn, Sungsoo Park, Jinkyoo Machine Learning Generative Flow Networks (GFlowNets) are amortized sampling methods that learn a distribution over discrete objects proportional to their rewards. GFlowNets exhibit a remarkable ability to generate diverse samples, yet occasionally struggle to consistently produce samples with high rewards due to over-exploration on wide sample space. This paper proposes to train GFlowNets with local search, which focuses on exploiting high-rewarded sample space to resolve this issue. Our main idea is to explore the local neighborhood via backtracking and reconstruction guided by backward and forward policies, respectively. This allows biasing the samples toward high-reward solutions, which is not possible for a typical GFlowNet solution generation scheme, which uses the forward policy to generate the solution from scratch. Extensive experiments demonstrate a remarkable performance improvement in several biochemical tasks. Source code is available: \url{https://github.com/dbsxodud-11/ls_gfn}.
title	Local Search GFlowNets
topic	Machine Learning
url	https://arxiv.org/abs/2310.02710