Programmatic RL is an upcoming field of research where explainable, adaptable and executable policies are created from black box models. We want to extend the current SOTA with a language- agnostic approach, first abstracting the policy before expressing it in a target language.
Deproost, S & Nowe, A 2024, 'Realizing executable programs from an abstract of the policy', Belgium-Netherlands Workshop on Reinforcement Learning 2024, Amsterdam, Netherlands, 10/06/24.
Deproost, S., & Nowe, A. (2024). Realizing executable programs from an abstract of the policy. Poster session presented at Belgium-Netherlands Workshop on Reinforcement Learning 2024, Amsterdam, Netherlands.
@conference{80f8b853341f453886b5540107d63df1,
title = "Realizing executable programs from an abstract of the policy",
abstract = "Programmatic RL is an upcoming field of research where explainable, adaptable and executable policies are created from black box models. We want to extend the current SOTA with a language- agnostic approach, first abstracting the policy before expressing it in a target language.",
keywords = "programmatic, reinfocement learning, explainability, explainable ai, XAI, XRL",
author = "Senne Deproost and Ann Nowe",
year = "2024",
month = jun,
day = "10",
language = "English",
note = "Belgium-Netherlands Workshop on Reinforcement Learning 2024, BeNeRL 2024 ; Conference date: 10-06-2024",
url = "https://www.benerl.org",
}