Omega-Regular Objectives in Model-Free Reinforcement Learning
- Submitting institution
-
The University of Liverpool
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 15836
- Type
- E - Conference contribution
- DOI
-
10.1007/978-3-030-17462-0_27
- Title of conference / published proceedings
- Tools and Algorithms for the Construction and Analysis of Systems. TACAS 2019. Lecture Notes in Computer Science
- First page
- 395
- Volume
- 11427
- Issue
- -
- ISSN
- 0302-9743
- Open access status
- Compliant
- Month of publication
- -
- Year of publication
- 2019
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
5
- Research group(s)
-
-
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper provides the first correct solution for model-free reinforcement learning with linear temporal logic objectives for Markov decision processes. It identifies errors in two previous attempts at this problem. The paper is part of a research programme on using advanced automata theory for model-free reinforcement learning. Further outputs that build on this one include "Model-Free Reinforcement Learning for Stochastic Parity Games" (CONCUR'20) and "Good-for-MDPs Automata for Probabilistic Analysis and Reinforcement Learning" (TACAS'20), both not REF returned.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -