1st International Workshop on Autonomous Network Management in 5G and Beyond Systems (ANSM 2022)

25 or 29 April, 2022


General Co-Chairs

TPC Co-Chairs


Miguel Camelo

University of Antwerp, IMEC



Danny De Vleeschauwer

Nokia Bell Labs



Francesc Wilhelmi

Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)


Technical Program Committee


Call For Papers

Beyond 5G (B5G) wireless systems are expected 1) to increase data rates significantly, 2) to provide ultra-low latency and enhanced connectivity of a massive number of devices, and 3) to bring improvements in network energy efficiency. Moreover, this new generation of networking systems aims to be fully Autonomous Networks (ANs) with management capabilities such as self-configuration, self-healing, self-optimizing, and self-evolving, aspects that today’s networks do not support as their management is largely manual with some automated assistance.

ANs provide a set of closed-loop systems that manage the resources within them, where each closed control loop can observe its environment and functionality thanks to novel cognitive capabilities, reason about those observations in the current situation, and take actions towards a set of well-defined goals. This process allows ANs to adjust their behaviour depending on the user needs and business goals when the environment changes, all of this with minimal human intervention. To successfully implement and deploy such ANs, significant novel contributions and innovations are required in the area of network management.

Advances in Artificial Intelligence (AI) during the last 5-10 years have provided a new set of algorithms and tools to solve challenging problems in multiple domains by learning complex relationships directly from data. These algorithms are also very promising to empower the new generation of intelligent decision engines that present self-dynamic capabilities to create smart business and network operations in a closed-loop fashion. However, new challenges emerge, such as: deciding when to use traditional management algorithms, AI-based models, or even hybrid approaches; design AI models tailor-made for network management problems; deciding where they have to be deployed (fog vs. edge); and how to manage their life cycle (from data harvesting to intelligent decision-making). Moreover, important questions arise regarding the usage of AI for communications, including sustainability (energy consumption and carbon footprint), reliability (ML performance in unseen situations), or responsiveness (real-time vs. non-real-time ML training). Part of this vision is embodied in the H2020-DAEMON project (https://h2020daemon.eu/), which develops and implements innovative and pragmatic approaches to Network Intelligence (NI) design that enable high performance, sustainable and extremely reliable zero-touch network system.

This workshop focuses on novel research in algorithms (especially AI-based) for the management of ANs in B5G systems. More specifically, we encourage original paper submissions, which have not been published or submitted elsewhere, from academia and industry presenting novel research on the most recent advances, frameworks, models, and approaches for management of autonomous networks using AI. We are also interested in articles revising the state-of-the-art of this topic, showing recent major advances and discoveries, significant gaps in the research, current standardization status, and new future issues.

Extended versions of the best paper(s) may be considered for fast-tracking to the Journal of Network and Systems Management (confirmed, https://www.springer.com/journal/10922, IF 2.026). The decision will depend on the quality and scope of the paper(s) and its (their) potential to spark lively discussion(s) at the workshop. The final decision will be made by the co-chairs after the workshop.

Topics of interest include, but are not limited to:

  1. AI-based and hybrid approaches (AI + non-AI) algorithms for AN management (closed loop control):
    • Decentralized management of multi-domain ANs.
    • End-to-end lifecycle of ANs.
    • Network resource and service automation and orchestration.
    • Automated network operation and maintenance.
    • Intelligent service provisioning and assurance.
    • Wireless network coverage optimization and assurance.
    • Wireless network energy saving.
    • Intelligent slice lifecycle management.
    • Metadata-driven policies to recognize and incorporate new knowledge in AN.
    • Efficient resource allocation and scheduling (e.g., spectrum, storage, computing, processing).
    • Network state prediction and forecasting for AN.
    • Network monitoring systems (traffic recognition, anomaly detection, etc.) for AN.
    • AN management in resource-constrained environments.
    • AN management in Mobile Radio Access Networks (RANs) and Wireless Local Area Networks (WLANs).
    • AN management in cognitive radio networks and opportunistic spectrum access.
    • AN management in Mobile Edge Computing (MEC).
    • AN Management in Transport Networks (TNs).
    • End-to-end management of AN.
    • Security provision in ANs
  2. Framework, algorithms, techniques, and tools to build and support AI-based solutions for ANs:
    • Architectures and frameworks to integrate AI in AN management (e.g., MLOps optimized for networking).
    • Algorithms, techniques, methods and tools from AI adapted to network management (e.g., online learning, reinforcement learning, hierarchical learning, multi-task learning, federated learning, transfer learning, spike neural networks, graph neural networks, resource-aware AI, explainable AI, one-shot learning, etc.).
    • Guidelines for designing and benchmarking AI-based solutions for network management.
    • Energy-efficient techniques for data pre-processing, resource allocation, and model training and deployment of AI-based algorithms.
    • Lightweight communication protocols for AI-embedded network management (e.g., neural networks compression).
    • AI-enabled network simulation tools, testbeds, or hardware implementations.
Scroll to Top