Transactions on Artificial Intelligence

Distributed Intelligence of Underwater Agents: Learning, Coordination, and Emergent Behaviors

Underwater environments are becoming increasingly important for ocean observation, marine resource exploration, environmental monitoring, offshore infrastructure inspection, disaster response, and autonomous scientific discovery. Recent advances in autonomous underwater vehicles, underwater sensor networks, marine robots, and multi-agent cyber-physical systems have created new opportunities for intelligent underwater missions that require groups of agents to perceive, learn, communicate, coordinate, and adapt in highly dynamic and uncertain environments. Unlike terrestrial and aerial domains, underwater systems face unique constraints, including severe attenuation of wireless signals, low-bandwidth and high-latency acoustic communications, GPS-denied localization, limited visibility, complex hydrodynamics, energy limitations, sparse supervision, and significant domain shifts between simulation and real-world deployment. These constraints make centralized intelligence difficult to deploy and call for distributed, adaptive, and communication-efficient artificial intelligence methods that can operate under partial observability, uncertainty, and intermittent connectivity. Distributed intelligence of underwater agents aims to enable multiple autonomous agents to learn from local observations, share knowledge efficiently, coordinate actions, and exhibit robust collective behaviors. This research direction spans multi-agent reinforcement learning, distributed optimization, federated and continual learning, graph learning, bio-inspired swarm intelligence, underwater perception, cooperative control, trustworthy AI, and emergent behavior analysis. It is also closely aligned with the broader goals of artificial intelligence: developing generalizable, explainable, safe, and autonomous learning systems capable of acting in complex real-world environments. This Call for Papers seeks high-quality original contributions that advance the theories, algorithms, architectures, and experimental validation of distributed intelligence for underwater agents. Submissions are expected to provide clear AI contributions, including new learning paradigms, coordination mechanisms, distributed decision-making frameworks, trustworthy autonomy methods, benchmark environments, or real-world demonstrations for underwater multi-agent systems. Topics of Interest Topics include, but are not limited to: 1. Distributed Learning for Underwater Multi-Agent Systems: Multi-agent reinforcement learning, federated learning, continual learning, self-supervised learning, and distributed optimization for autonomous underwater agents operating under limited communication, sparse supervision, and dynamic environments. 2. Coordination and Decision-Making under Underwater Constraints: Decentralized planning, co- operative control, task allocation, formation control, consensus, and decision-making under partial observability, delayed communication, uncertain localization, and energy constraints. 3. Emergent Behaviors and Swarm Intelligence: Learning, modeling, analysis, and interpretation of collective intelligence, self-organization, bio-inspired underwater swarms, robust group behaviors, and adaptive cooperation among heterogeneous underwater agents. 4. Intelligent Communication among Underwater Agents: Agent-to-agent communication protocols, semantic communication, goal-oriented information exchange, learned communication languages, intent sharing, negotiation, and interaction mechanisms for distributed underwater intelligence. 5. Communication-Efficient Intelligence and Knowledge Sharing: Learning and coordination under low-bandwidth, intermittent, and high-latency underwater communication links, including knowledge distillation, compressed learning, adaptive message passing, and communication-aware multi- agent policies. 6. Perception, Localization, and World Modeling for Underwater Agents: AI-enabled underwater perception, sonar and vision-based sensing, multi-modal fusion, cooperative localization, mapping, world models, and uncertainty-aware environmental understanding for distributed underwater autonomy. 7. Robust, Safe, and Explainable Underwater AI: Robust learning, sim-to-real transfer, domain adaptation, uncertainty quantification, explainable decision-making, safety-aware learning, fault tolerance, and trustworthy autonomy in complex underwater environments. 8. Benchmarks, Simulators, and Field Validation: Datasets, simulation platforms, digital twins, re- producible benchmarks, evaluation protocols, and real-world demonstrations for learning, coordination, and emergent behaviors of underwater agent systems. Authors are invited to submit original and unpublished manuscripts that are not currently under review elsewhere. Manuscripts should follow the formatting and submission requirements of T ransactions on Artificial Intelligence . All submissions will undergo the journal’s standard peer-review process. Submissions should clearly indicate their relevance to the topic on “Distributed Intelligence of Underwater Agents: Learning, Coordination, and Emergent Behaviors.” When submitting, authors should select the corresponding category in the journal’s submission system, if available. Submissions are expected to make a clear contribution to artificial intelligence, such as new learning algorithms, distributed decision-making methods, theoretical analysis, trustworthy AI frameworks, benchmark environments, or experimentally validated intelligent systems. Papers that only report routine engineering applications without sufficient AI novelty may be considered out of scope. Academic Editors Dr. Yusha Liu (yusha.liu@uestc.edu.cn), University of Electronic Science and Technology of China, China Dr. Jingxuan Chen (chenjingxuan@buaa.edu.cn), Hong Kong University of Science and Technology, Hong Kong Submission Guidelines To submit your manuscript, please go to the Transactions on Artificial Intelligence journal website here and follow the procedures for manuscript submission. Author Submission Guidelines can be found at: Instruction for Authors. There is no Article Processing Charge for all submissions and accepted papers. Important Dates Manuscript Submission Deadline: 31 August 2026 First-Round Review Notification: 15 October 2026 Revised Manuscripts Due: 31 November2026 Final Decision: 31 December 2026 Expected Publication: January 2027 All manuscripts will be peer-reviewed following the established policies and procedures of the journal. The final papers will be selected for publication depending on the results of the peer-review process and the reviews of the Academic Editors and Editor-in-Chief.

Scilight Press

Transactions on Artificial Intelligence

Articles

Announcements