Salvato in:
Dettagli Bibliografici
Autori principali: Preuett III, Larry, Zhang, Qiuyi, Ahmad, Muhammad Aurangzeb
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2505.06518
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914532854595584
author Preuett III, Larry
Zhang, Qiuyi
Ahmad, Muhammad Aurangzeb
author_facet Preuett III, Larry
Zhang, Qiuyi
Ahmad, Muhammad Aurangzeb
contents In many real-world planning tasks, agents must tackle uncertainty about the environment's state and variability in the outcomes induced by stochastic dynamics and rewards. Motivated by recent progress in world model approaches, where latent models approximate beliefs and support planning, we extend Distributional Reinforcement Learning (DistRL), which models the entire return distribution for fully observable domains, to Partially Observable Markov Decision Processes (POMDPs). Concretely, we introduce new distributional Bellman operators for partial observability and prove their convergence under the supremum p-Wasserstein metric. We also propose a finite representation of these return distributions via psi-vectors, generalizing the classical alpha-vectors in POMDP solvers. Building on this, we develop Distributional Point-Based Value Iteration (DPBVI), which integrates psi-vectors into a standard point-based backup procedure, bridging DistRL and POMDP planning. Our experiments demonstrate that DPBVI recovers classical Point-Based Value Iteration (PBVI) in the risk-neutral case, validating the distributional extension.
format Preprint
id arxiv_https___arxiv_org_abs_2505_06518
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Provable Distributional Value Iteration under Partial Observability
Preuett III, Larry
Zhang, Qiuyi
Ahmad, Muhammad Aurangzeb
Artificial Intelligence
In many real-world planning tasks, agents must tackle uncertainty about the environment's state and variability in the outcomes induced by stochastic dynamics and rewards. Motivated by recent progress in world model approaches, where latent models approximate beliefs and support planning, we extend Distributional Reinforcement Learning (DistRL), which models the entire return distribution for fully observable domains, to Partially Observable Markov Decision Processes (POMDPs). Concretely, we introduce new distributional Bellman operators for partial observability and prove their convergence under the supremum p-Wasserstein metric. We also propose a finite representation of these return distributions via psi-vectors, generalizing the classical alpha-vectors in POMDP solvers. Building on this, we develop Distributional Point-Based Value Iteration (DPBVI), which integrates psi-vectors into a standard point-based backup procedure, bridging DistRL and POMDP planning. Our experiments demonstrate that DPBVI recovers classical Point-Based Value Iteration (PBVI) in the risk-neutral case, validating the distributional extension.
title Provable Distributional Value Iteration under Partial Observability
topic Artificial Intelligence
url https://arxiv.org/abs/2505.06518