Explainable Reinforcement and Causal Learning for Improving Trust to 6G Stakeholders

Miguel Arana-Catania; Amir Sonee; Abdul Manan Khan; Kavan Fatehi; Yun Tang; Bailu Jin; Anna Soligo; David Boyle; Radu Calinescu; Poonam Yadav; Hamed Ahmadi; Antonios Tsourdos; Weisi Guo; Alessandra Russo

doi:10.1109/OJCOMS.2025.3563415

Explainable Reinforcement and Causal Learning for Improving Trust to 6G Stakeholders

Miguel Arana-Catania, Amir Sonee, Abdul Manan Khan, Kavan Fatehi, Yun Tang, Bailu Jin, Anna Soligo, David Boyle, Radu Calinescu, Poonam Yadav, Hamed Ahmadi, Antonios Tsourdos, Weisi Guo^*, Alessandra Russo

^*Corresponding author for this work

Research output: Contribution to journal › Review article › peer-review

Abstract

Future telecommunications will increasingly integrate AI capabilities into network infrastructures to deliver seamless and harmonized services closer to end-users. However, this progress also raises significant trust and safety concerns. The machine learning systems orchestrating these advanced services will widely rely on deep reinforcement learning (DRL) to process multi-modal requirements datasets and make semantically modulated decisions, introducing three major challenges: (1) First, we acknowledge that most explainable AI research is stakeholder-agnostic while, in reality, the explanations must cater for diverse telecommunications stakeholders, including network service providers, legal authorities, and end users, each with unique goals and operational practices; (2) Second, DRL lacks prior models or established frameworks to guide the creation of meaningful long-term explanations of the agent's behaviour in a goal-oriented RL task, and we introduce state-of-the-art approaches such as reward machine and sub-goal automata that can be universally represented and easily manipulated by logic programs and verifiably learned by inductive logic programming of answer set programs; (3) Third, most explainability approaches focus on correlation rather than causation, and we emphasise that understanding causal learning can further enhance 6G network optimisation. Together, in our judgement they form crucial enabling technologies for trustworthy services in 6G. This review offers a timely resource for academic researchers and industry practitioners by highlighting the methodological advancements needed for explainable DRL (X-DRL) in 6G. It identifies key stakeholder groups, maps their needs to X-DRL solutions, and presents case studies showcasing practical applications. By identifying and analysing these challenges in the context of 6G case studies, this work aims to inform future research, transform industry practices, and highlight unresolved gaps in this rapidly evolving field.

Original language	English
Pages (from-to)	4101-4125
Number of pages	25
Journal	IEEE Open Journal of the Communications Society
Volume	6
DOIs	https://doi.org/10.1109/OJCOMS.2025.3563415
Publication status	Published - 22 Apr 2025

Bibliographical note

Publisher Copyright:
© 2020 IEEE.

Keywords

6G
causal learning
explainable AI
reinforcement learning
stakeholders
trust

Access to Document

10.1109/OJCOMS.2025.3563415Licence: CC BY

Explainable Reinforcement and Causal Learning for Improving Trust to 6G Stakeholders
© 2025 The Authors
Final published version, 4.65 MBLicence: CC BY

Cite this

Arana-Catania, M., Sonee, A., Khan, A. M., Fatehi, K., Tang, Y., Jin, B., Soligo, A., Boyle, D., Calinescu, R., Yadav, P., Ahmadi, H., Tsourdos, A., Guo, W., & Russo, A. (2025). Explainable Reinforcement and Causal Learning for Improving Trust to 6G Stakeholders. IEEE Open Journal of the Communications Society, 6, 4101-4125. https://doi.org/10.1109/OJCOMS.2025.3563415

@article{bf23e2d806cc4f65b1f26c7b996bcadf,

title = "Explainable Reinforcement and Causal Learning for Improving Trust to 6G Stakeholders",

abstract = "Future telecommunications will increasingly integrate AI capabilities into network infrastructures to deliver seamless and harmonized services closer to end-users. However, this progress also raises significant trust and safety concerns. The machine learning systems orchestrating these advanced services will widely rely on deep reinforcement learning (DRL) to process multi-modal requirements datasets and make semantically modulated decisions, introducing three major challenges: (1) First, we acknowledge that most explainable AI research is stakeholder-agnostic while, in reality, the explanations must cater for diverse telecommunications stakeholders, including network service providers, legal authorities, and end users, each with unique goals and operational practices; (2) Second, DRL lacks prior models or established frameworks to guide the creation of meaningful long-term explanations of the agent's behaviour in a goal-oriented RL task, and we introduce state-of-the-art approaches such as reward machine and sub-goal automata that can be universally represented and easily manipulated by logic programs and verifiably learned by inductive logic programming of answer set programs; (3) Third, most explainability approaches focus on correlation rather than causation, and we emphasise that understanding causal learning can further enhance 6G network optimisation. Together, in our judgement they form crucial enabling technologies for trustworthy services in 6G. This review offers a timely resource for academic researchers and industry practitioners by highlighting the methodological advancements needed for explainable DRL (X-DRL) in 6G. It identifies key stakeholder groups, maps their needs to X-DRL solutions, and presents case studies showcasing practical applications. By identifying and analysing these challenges in the context of 6G case studies, this work aims to inform future research, transform industry practices, and highlight unresolved gaps in this rapidly evolving field.",

keywords = "6G, causal learning, explainable AI, reinforcement learning, stakeholders, trust",

author = "Miguel Arana-Catania and Amir Sonee and Khan, {Abdul Manan} and Kavan Fatehi and Yun Tang and Bailu Jin and Anna Soligo and David Boyle and Radu Calinescu and Poonam Yadav and Hamed Ahmadi and Antonios Tsourdos and Weisi Guo and Alessandra Russo",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.",

year = "2025",

month = apr,

day = "22",

doi = "10.1109/OJCOMS.2025.3563415",

language = "English",

volume = "6",

pages = "4101--4125",

journal = "IEEE Open Journal of the Communications Society",

issn = "2644-125X",

publisher = "IEEE",

}

TY - JOUR

T1 - Explainable Reinforcement and Causal Learning for Improving Trust to 6G Stakeholders

AU - Arana-Catania, Miguel

AU - Sonee, Amir

AU - Khan, Abdul Manan

AU - Fatehi, Kavan

AU - Tang, Yun

AU - Jin, Bailu

AU - Soligo, Anna

AU - Boyle, David

AU - Calinescu, Radu

AU - Yadav, Poonam

AU - Ahmadi, Hamed

AU - Tsourdos, Antonios

AU - Guo, Weisi

AU - Russo, Alessandra

PY - 2025/4/22

Y1 - 2025/4/22

N2 - Future telecommunications will increasingly integrate AI capabilities into network infrastructures to deliver seamless and harmonized services closer to end-users. However, this progress also raises significant trust and safety concerns. The machine learning systems orchestrating these advanced services will widely rely on deep reinforcement learning (DRL) to process multi-modal requirements datasets and make semantically modulated decisions, introducing three major challenges: (1) First, we acknowledge that most explainable AI research is stakeholder-agnostic while, in reality, the explanations must cater for diverse telecommunications stakeholders, including network service providers, legal authorities, and end users, each with unique goals and operational practices; (2) Second, DRL lacks prior models or established frameworks to guide the creation of meaningful long-term explanations of the agent's behaviour in a goal-oriented RL task, and we introduce state-of-the-art approaches such as reward machine and sub-goal automata that can be universally represented and easily manipulated by logic programs and verifiably learned by inductive logic programming of answer set programs; (3) Third, most explainability approaches focus on correlation rather than causation, and we emphasise that understanding causal learning can further enhance 6G network optimisation. Together, in our judgement they form crucial enabling technologies for trustworthy services in 6G. This review offers a timely resource for academic researchers and industry practitioners by highlighting the methodological advancements needed for explainable DRL (X-DRL) in 6G. It identifies key stakeholder groups, maps their needs to X-DRL solutions, and presents case studies showcasing practical applications. By identifying and analysing these challenges in the context of 6G case studies, this work aims to inform future research, transform industry practices, and highlight unresolved gaps in this rapidly evolving field.

AB - Future telecommunications will increasingly integrate AI capabilities into network infrastructures to deliver seamless and harmonized services closer to end-users. However, this progress also raises significant trust and safety concerns. The machine learning systems orchestrating these advanced services will widely rely on deep reinforcement learning (DRL) to process multi-modal requirements datasets and make semantically modulated decisions, introducing three major challenges: (1) First, we acknowledge that most explainable AI research is stakeholder-agnostic while, in reality, the explanations must cater for diverse telecommunications stakeholders, including network service providers, legal authorities, and end users, each with unique goals and operational practices; (2) Second, DRL lacks prior models or established frameworks to guide the creation of meaningful long-term explanations of the agent's behaviour in a goal-oriented RL task, and we introduce state-of-the-art approaches such as reward machine and sub-goal automata that can be universally represented and easily manipulated by logic programs and verifiably learned by inductive logic programming of answer set programs; (3) Third, most explainability approaches focus on correlation rather than causation, and we emphasise that understanding causal learning can further enhance 6G network optimisation. Together, in our judgement they form crucial enabling technologies for trustworthy services in 6G. This review offers a timely resource for academic researchers and industry practitioners by highlighting the methodological advancements needed for explainable DRL (X-DRL) in 6G. It identifies key stakeholder groups, maps their needs to X-DRL solutions, and presents case studies showcasing practical applications. By identifying and analysing these challenges in the context of 6G case studies, this work aims to inform future research, transform industry practices, and highlight unresolved gaps in this rapidly evolving field.

KW - 6G

KW - causal learning

KW - explainable AI

KW - reinforcement learning

KW - stakeholders

KW - trust

UR - http://www.scopus.com/inward/record.url?scp=105003489540&partnerID=8YFLogxK

U2 - 10.1109/OJCOMS.2025.3563415

DO - 10.1109/OJCOMS.2025.3563415

M3 - Review article

AN - SCOPUS:105003489540

SN - 2644-125X

VL - 6

SP - 4101

EP - 4125

JO - IEEE Open Journal of the Communications Society

JF - IEEE Open Journal of the Communications Society

ER -

Explainable Reinforcement and Causal Learning for Improving Trust to 6G Stakeholders

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Cite this