Utilizing Explainability Techniques for Reinforcement Learning Model Assurance

Alexander Tapley, Kyle Gatesman, Luis Robaina, Brett Bissey, Joseph Weissman
11/27/2023
cs.LGcs.AI

Official code: mitre/arlin (11 stars, last commit 2024-02-06).

Abstract

Explainable Reinforcement Learning (XRL) can provide transparency into the decision-making process of a Deep Reinforcement Learning (DRL) model and increase user trust and adoption in real-world use cases. By utilizing XRL techniques, researchers can identify potential vulnerabilities within a trained DRL model prior to deployment, therefore limiting the potential for mission failure or mistakes by the system. This paper introduces the ARLIN (Assured RL Model Interrogation) Toolkit, an open-source Python library that identifies potential vulnerabilities and critical points within trained DRL models through detailed, human-interpretable explainability outputs. To illustrate ARLIN's effectiveness, we provide explainability visualizations and vulnerability analysis for a publicly available DRL model. The open-source code repository is available for download at https://github.com/mitre/arlin.

View on arXivView PDF

Code Implementations(1)

mitre/arlinOfficial100%

ARLIN is a research library written in Python that provides explainability outputs and vulnerability detection for DRL models, specifically designed to increase model assurance and identify vulnerabilities and potential points of failure within a trained model. https://mitre.github.io/arlin/

111PythonNov 8, 20232 years agoNOASSERTION
adversarial-machine-learningexplainable-aipythonreinforcement-learning

Similar Papers

A Survey on Explainable Deep Reinforcement Learning

86% match2025

xSRL: Safety-Aware Explainable Reinforcement Learning -- Safety as a Product of Explainability

86% match2024

Explainable Deep Reinforcement Learning: State of the Art and Challenges

83% match2023

SafeRL-Lite: A Lightweight, Explainable, and Constrained Reinforcement Learning Library

82% match1 repo32025

From Explainability to Interpretability: Interpretable Policies in Reinforcement Learning Via Model Explanation

82% match2025

Textual Explanations and Their Evaluations for Reinforcement Learning Policy

82% match2026

Cite this paper

@article{tapley2023utilizing,
  title  = {Utilizing Explainability Techniques for Reinforcement Learning Model Assurance},
  author = {Alexander Tapley and Kyle Gatesman and Luis Robaina and Brett Bissey and Joseph Weissman},
  year   = {2023},
  eprint = {2311.15838},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2311.15838v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2311.15838.svg)](https://codeofpaper.com/paper/2311.15838)

Discussion