Partially Observable Stochastic Games with Neural Perception Mechanisms

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yan, Rui, Santos, Gabriel, Norman, Gethin, Parker, David, Kwiatkowska, Marta
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911937079541760
author Yan, Rui
Santos, Gabriel
Norman, Gethin
Parker, David
Kwiatkowska, Marta
author_facet Yan, Rui
Santos, Gabriel
Norman, Gethin
Parker, David
Kwiatkowska, Marta
contents Stochastic games are a well established model for multi-agent sequential decision making under uncertainty. In practical applications, though, agents often have only partial observability of their environment. Furthermore, agents increasingly perceive their environment using data-driven approaches such as neural networks trained on continuous data. We propose the model of neuro-symbolic partially-observable stochastic games (NS-POSGs), a variant of continuous-space concurrent stochastic games that explicitly incorporates neural perception mechanisms. We focus on a one-sided setting with a partially-informed agent using discrete, data-driven observations and another, fully-informed agent. We present a new method, called one-sided NS-HSVI, for approximate solution of one-sided NS-POSGs, which exploits the piecewise constant structure of the model. Using neural network pre-image analysis to construct finite polyhedral representations and particle-based representations for beliefs, we implement our approach and illustrate its practical applicability to the analysis of pedestrian-vehicle and pursuit-evasion scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2310_11566
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Partially Observable Stochastic Games with Neural Perception Mechanisms
Yan, Rui
Santos, Gabriel
Norman, Gethin
Parker, David
Kwiatkowska, Marta
Computer Science and Game Theory
Machine Learning
Stochastic games are a well established model for multi-agent sequential decision making under uncertainty. In practical applications, though, agents often have only partial observability of their environment. Furthermore, agents increasingly perceive their environment using data-driven approaches such as neural networks trained on continuous data. We propose the model of neuro-symbolic partially-observable stochastic games (NS-POSGs), a variant of continuous-space concurrent stochastic games that explicitly incorporates neural perception mechanisms. We focus on a one-sided setting with a partially-informed agent using discrete, data-driven observations and another, fully-informed agent. We present a new method, called one-sided NS-HSVI, for approximate solution of one-sided NS-POSGs, which exploits the piecewise constant structure of the model. Using neural network pre-image analysis to construct finite polyhedral representations and particle-based representations for beliefs, we implement our approach and illustrate its practical applicability to the analysis of pedestrian-vehicle and pursuit-evasion scenarios.
title Partially Observable Stochastic Games with Neural Perception Mechanisms
topic Computer Science and Game Theory
Machine Learning
url https://arxiv.org/abs/2310.11566