Scenario-guided temporal prototypes in reinforcement learning

Dobravec, Blaž (author), Žabkar, Jure (author)

Deep reinforcement learning policies are hard to deploy in safety-critical settings, because they fail to explain why a sequence of actions is taken. We introduce an intrinsically interpretable framework that learns compact summaries of recurring behavior and uses them for case-based decision making. Our method (i) discovers global regimes by grouping trajectories into a small set of recurrent patterns and (ii) learns a prototype-conditioned local policy that maps the current short-horizon pattern to an action (“this matches prototype X → take action Y”). Each action is accompanied by a similarity score to relevant prototypes, which provide the explanations. We evaluate our approach on two domains: (1) CarRacing (pixel-based continuous control) and (2) a real voltage-control problem in low-voltage distribution networks. Our results indicate that the method provides clear pre hoc explanations while keeping task performance close to the reference policy.

• Language: English
• Keywords: deep reinforcement learning, explainability, prototypes
• Work type: Article
• Typology: 1.01 - Original Scientific Article
• Organization: FRI - Faculty of Computer and Information Science
• Publication status: Published
• Publication version: Version of Record
• Year: 2026
• Number of pages: 14 str.
• Numbering: Vol. 8, iss. 1, art. 21
• PID: 20.500.12556/RUL-181718
• UDC: 004.8:007.52
• ISSN on article: 2504-4990
• DOI: 10.3390/make8010021
• COBISS.SI-ID: 265133315
• Publication date in RUL: 14.04.2026

Record is a part of a journal

• Title: Machine learning and knowledge extraction
• Publisher: MDPI
• ISSN: 2504-4990
• COBISS.SI-ID: 1537706179

Licences

• License: CC BY 4.0, Creative Commons Attribution 4.0 International
• Link: http://creativecommons.org/licenses/by/4.0/
• Description: This is the standard Creative Commons license that gives others maximum freedom to do what they want with the work as long as they credit the author.

Secondary language

• Language: Slovenian
• Keywords: globoko spodbujevano učenje, razložljivost, prototipi

Projects

• Funder: ARIS - Slovenian Research and Innovation Agency
• Project number: L2-4436
• Name: Optimizacija obratovanja nizkonapetostnega distribucijskega omrežja z integrirano fleksibilnostjo v realnem času s pomočjo globokega spodbujevanega učenja (DRIFT)

• Funder: ARIS - Slovenian Research and Innovation Agency
• Project number: P2-0209
• Name: Umetna inteligenca in inteligentni sistemi