Theory and Practice of Safe and Secure Learning in Agentic AI

Dear Colleagues,

The rapidly evolving fields of Machine Learning (ML) and Multi-Agent Systems (MASs) are reshaping intelligent decision-making in complex, distributed, and interactive environments. As autonomous agents become increasingly embedded in real-world infrastructures, from transportation and energy systems to finance, cybersecurity, and robotics, ensuring both theoretical rigor and system-level safety have become critically important.

This Special Issue focuses on foundational advances in ML for MAS, with a particular emphasis on safety, security, robustness, and trustworthy decision-making. It aims to bring together cutting-edge research that bridges theoretical developments with practical challenges arising in multi-agent learning systems operating under uncertainty, adversarial conditions, and partial observability.

We welcome contributions that develop new mathematical and algorithmic foundations for learning in multi-agent settings, including, but not limited to, convergence analysis, stability guarantees, equilibrium learning, and scalable coordination mechanisms. In parallel, we strongly encourage submissions addressing safety-critical aspects such as adversarial robustness, secure communication, resilient coordination, fault tolerance, and formal verification of learned multi-agent policies.

In addition to theoretical contributions, the Special Issue seeks applied and empirical studies demonstrating how safe and secure ML-MAS techniques can be deployed in real-world domains. These include autonomous vehicle fleets, swarm robotics, smart cities, critical infrastructure protection, distributed cyber defence, and socio-technical systems where multiple learning agents interact under strategic and potentially adversarial settings.

By integrating foundational theory with safety and security considerations, this Special Issue aims to advance the development of reliable, transparent, and resilient multi-agent learning systems capable of operating in complex real-world environments.

Topics of interest include (but are not limited to):

• Foundational theory of machine learning in multi-agent systems.
• Game-theoretic learning and equilibrium analysis in MAS.
• Safe reinforcement learning in multi-agent environments.
• Secure and adversarial robust multi-agent learning.
• Decentralized and distributed learning with formal guarantees.
• Stability, convergence, and optimality in multi-agent learning dynamics.
• Fault-tolerant and resilient coordination mechanisms.
• Privacy-preserving and secure communication in MAS.
• Testing, Verification, validation, and interpretability of multi-agent policies.
• Trustworthy AI in cooperative and competitive multi-agent settings.
• Adversarial attacks and defences in multi-agent learning systems.
• Real-world applications in robotics, autonomous systems, cybersecurity, and critical infrastructure.

Dr. Abdur Rakib
Dr. Mehmet Aydin
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-anonymized peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Applied Sciences is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• foundational theory of machine learning in multi-agent systems
• game-theoretic learning and equilibrium analysis in MAS
• safe reinforcement learning in multi-agent environments
• secure and adversarial robust multi-agent learning
• decentralized and distributed learning with formal guarantees
• stability, convergence, and optimality in multi-agent learning dynamics
• fault-tolerant and resilient coordination mechanisms
• privacy-preserving and secure communication in MAS
• testing, verification, validation, and interpretability of multi-agent policies
• trustworthy AI in cooperative and competitive multi-agent settings
• adversarial attacks and defences in multi-agent learning systems
• real-world applications in robotics, autonomous systems