Bifurcated Generative Flow Networks

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Li, Chunhui, Liu, Cheng-Hao, Liu, Dianbo, Cai, Qingpeng, Pan, Ling
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911902132600832
author Li, Chunhui
Liu, Cheng-Hao
Liu, Dianbo
Cai, Qingpeng
Pan, Ling
author_facet Li, Chunhui
Liu, Cheng-Hao
Liu, Dianbo
Cai, Qingpeng
Pan, Ling
contents Generative Flow Networks (GFlowNets), a new family of probabilistic samplers, have recently emerged as a promising framework for learning stochastic policies that generate high-quality and diverse objects proportionally to their rewards. However, existing GFlowNets often suffer from low data efficiency due to the direct parameterization of edge flows or reliance on backward policies that may struggle to scale up to large action spaces. In this paper, we introduce Bifurcated GFlowNets (BN), a novel approach that employs a bifurcated architecture to factorize the flows into separate representations for state flows and edge-based flow allocation. This factorization enables BN to learn more efficiently from data and better handle large-scale problems while maintaining the convergence guarantee. Through extensive experiments on standard evaluation benchmarks, we demonstrate that BN significantly improves learning efficiency and effectiveness compared to strong baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2406_01901
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Bifurcated Generative Flow Networks
Li, Chunhui
Liu, Cheng-Hao
Liu, Dianbo
Cai, Qingpeng
Pan, Ling
Machine Learning
Generative Flow Networks (GFlowNets), a new family of probabilistic samplers, have recently emerged as a promising framework for learning stochastic policies that generate high-quality and diverse objects proportionally to their rewards. However, existing GFlowNets often suffer from low data efficiency due to the direct parameterization of edge flows or reliance on backward policies that may struggle to scale up to large action spaces. In this paper, we introduce Bifurcated GFlowNets (BN), a novel approach that employs a bifurcated architecture to factorize the flows into separate representations for state flows and edge-based flow allocation. This factorization enables BN to learn more efficiently from data and better handle large-scale problems while maintaining the convergence guarantee. Through extensive experiments on standard evaluation benchmarks, we demonstrate that BN significantly improves learning efficiency and effectiveness compared to strong baselines.
title Bifurcated Generative Flow Networks
topic Machine Learning
url https://arxiv.org/abs/2406.01901