Maximally Permissive Reward Machines

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Varricchione, Giovanni, Alechina, Natasha, Dastani, Mehdi, Logan, Brian
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909287963426816
author Varricchione, Giovanni
Alechina, Natasha
Dastani, Mehdi
Logan, Brian
author_facet Varricchione, Giovanni
Alechina, Natasha
Dastani, Mehdi
Logan, Brian
contents Reward machines allow the definition of rewards for temporally extended tasks and behaviors. Specifying "informative" reward machines can be challenging. One way to address this is to generate reward machines from a high-level abstract description of the learning environment, using techniques such as AI planning. However, previous planning-based approaches generate a reward machine based on a single (sequential or partial-order) plan, and do not allow maximum flexibility to the learning agent. In this paper we propose a new approach to synthesising reward machines which is based on the set of partial order plans for a goal. We prove that learning using such "maximally permissive" reward machines results in higher rewards than learning using RMs based on a single plan. We present experimental results which support our theoretical claims by showing that our approach obtains higher rewards than the single-plan approach in practice.
format Preprint
id arxiv_https___arxiv_org_abs_2408_08059
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Maximally Permissive Reward Machines
Varricchione, Giovanni
Alechina, Natasha
Dastani, Mehdi
Logan, Brian
Machine Learning
Artificial Intelligence
68T05
Reward machines allow the definition of rewards for temporally extended tasks and behaviors. Specifying "informative" reward machines can be challenging. One way to address this is to generate reward machines from a high-level abstract description of the learning environment, using techniques such as AI planning. However, previous planning-based approaches generate a reward machine based on a single (sequential or partial-order) plan, and do not allow maximum flexibility to the learning agent. In this paper we propose a new approach to synthesising reward machines which is based on the set of partial order plans for a goal. We prove that learning using such "maximally permissive" reward machines results in higher rewards than learning using RMs based on a single plan. We present experimental results which support our theoretical claims by showing that our approach obtains higher rewards than the single-plan approach in practice.
title Maximally Permissive Reward Machines
topic Machine Learning
Artificial Intelligence
68T05
url https://arxiv.org/abs/2408.08059