Random Policy Evaluation Uncovers Policies of Generative Flow Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Haoran, Bengio, Emmanuel, Cai, Qingpeng, Pan, Ling
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909630680006656
author He, Haoran
Bengio, Emmanuel
Cai, Qingpeng
Pan, Ling
author_facet He, Haoran
Bengio, Emmanuel
Cai, Qingpeng
Pan, Ling
contents The Generative Flow Network (GFlowNet) is a probabilistic framework in which an agent learns a stochastic policy and flow functions to sample objects proportionally to an unnormalized reward function. A number of recent works explored connections between GFlowNets and maximum entropy (MaxEnt) RL, which modifies the standard objective of RL agents by learning an entropy-regularized objective. However, the relationship between GFlowNets and standard RL remains largely unexplored, despite the inherent similarities in their sequential decision-making nature. While GFlowNets can discover diverse solutions through specialized flow-matching objectives, connecting them can simplify their implementation through established RL principles and improve RL's diverse solution discovery capabilities. In this paper, we bridge this gap by revealing a fundamental connection between GFlowNets and one RL's most basic components -- policy evaluation. Surprisingly, we find that the value function obtained from evaluating a uniform policy is closely associated with the flow functions in GFlowNets through the lens of flow iteration under certain structural conditions. Building upon these insights, we introduce a rectified random policy evaluation (RPE) algorithm, which achieves the same reward-matching effect as GFlowNets based on simply evaluating a fixed random policy in these cases, offering a new perspective. Empirical results across extensive benchmarks demonstrate that RPE achieves competitive results compared to previous approaches, shedding light on the previously overlooked connection between (non-MaxEnt) RL and GFlowNets.
format Preprint
id arxiv_https___arxiv_org_abs_2406_02213
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Random Policy Evaluation Uncovers Policies of Generative Flow Networks
He, Haoran
Bengio, Emmanuel
Cai, Qingpeng
Pan, Ling
Machine Learning
The Generative Flow Network (GFlowNet) is a probabilistic framework in which an agent learns a stochastic policy and flow functions to sample objects proportionally to an unnormalized reward function. A number of recent works explored connections between GFlowNets and maximum entropy (MaxEnt) RL, which modifies the standard objective of RL agents by learning an entropy-regularized objective. However, the relationship between GFlowNets and standard RL remains largely unexplored, despite the inherent similarities in their sequential decision-making nature. While GFlowNets can discover diverse solutions through specialized flow-matching objectives, connecting them can simplify their implementation through established RL principles and improve RL's diverse solution discovery capabilities. In this paper, we bridge this gap by revealing a fundamental connection between GFlowNets and one RL's most basic components -- policy evaluation. Surprisingly, we find that the value function obtained from evaluating a uniform policy is closely associated with the flow functions in GFlowNets through the lens of flow iteration under certain structural conditions. Building upon these insights, we introduce a rectified random policy evaluation (RPE) algorithm, which achieves the same reward-matching effect as GFlowNets based on simply evaluating a fixed random policy in these cases, offering a new perspective. Empirical results across extensive benchmarks demonstrate that RPE achieves competitive results compared to previous approaches, shedding light on the previously overlooked connection between (non-MaxEnt) RL and GFlowNets.
title Random Policy Evaluation Uncovers Policies of Generative Flow Networks
topic Machine Learning
url https://arxiv.org/abs/2406.02213