Artificial Intelligence (AI) and Machine Learning (ML) are transforming communication networks into intelligent, adaptive and autonomous systems. By analyzing massive volumes of network data, AI algorithms can identify traffic trends, detect specific patterns or deviations and anticipate potential events, such as network failures or congestion, before they occur. This enables proactive optimization of network performance, dynamic resource orchestration and automated fault management. ML-driven insights also strengthen cybersecurity by identifying anomalies and mitigating threats in real time or even in advance. As networks grow in complexity with 5G/6G, IoT, and cloud services, AI enables optimal resource utilization while ensuring resilience and maximizing performance and efficiency; thereby reducing the need for human intervention. The result is a self-healing, secure, and highly responsive infrastructure that meets the demands of modern digital ecosystems.

The modern communications landscape places unprecedented demands on telecommunication networks. They must support applications that require both ultra-low latency and high data rates, such as those related to extended reality. It must also scale to high-density user environments such as massive IoT deployments and connected vehicle ecosystems. At the same time, operators must manage heterogeneous networks that combine a wide variety of end-user devices and networking equipment. Additionally, mission-critical applications impose stringent reliability and security requirements that go beyond traditional policies and basic encryption mechanisms.

Current systems are based on traditional data processing flows that cannot handle real-time decision-making or massive data processing requirements, which renders them obsolete. AI/ML surpasses traditional data processing in its ability to predict or detect events, identify patterns, and solve complex allocation problems, particularly in scenarios involving massive volumes of data.

The research community has identified integrating AI/ML into communication networks as essential for addressing the challenges posed by new applications and services. This includes providing direct access to AI/ML applications at the edge and offering an intelligent communication substrate that natively integrates AI/ML into network operations. The goal is to improve functions such as management and orchestration, monitoring, security, packet processing, and others by adding intelligence.

This Research Topic encompasses a variety of research activities that leverage AI/ML to improve the operation and performance of communication networks, including:

- Event detection and prediction. Use AI/ML to predict congestion, detect anomalies, foresee network failures, and identify traffic patterns linked to attacks.

- Deep packet inspection and security. Analyze traffic flows for granular details (source, destination, protocol) to define and enforce security policies across devices.

- Traffic monitoring solutions. Implement efficient data collection and dynamic selection of relevant telemetry (e.g., In-Network Telemetry) to reduce costs and adapt to scenarios.

- Optimal path computation and traffic engineering. Compute routes based on traffic predictions, optimize TE configurations, and secure paths against interception or misdirection.

- AI-based Management and Orchestration (M&O). Enable intelligent resource allocation, mobility-aware handovers, self-healing, and intent-based operations for resilient 6G networks.

- AI and network programmability. Apply AI in softwarized and programmable networks, leveraging hardware accelerators (DPUs, NPUs, GPUs, P4 ASICs) for distributed AI tasks.

- Testbeds using real network equipment or processing devices will be of special interest.

Are you interested in contributing a manuscript to this Research Topic?

AI-Driven Operations in Communication Networks