This page lists all open topics in the area of AI/ML for communication networks. Interested students can contact the supervisor(s) to ask questions or discuss the details of a topic. It is also possible to propose your own topic in the area. An overview of AI/ML applications in networking can be found in [1].
Supervisors of our AI/ML work group: Christian Vielhaus, Johannes Busch, Vincent Latzko
Supervisor: Christian Vielhaus
Predicting network traffic facilitates various management and orchestration problems in communication networks. Traffic volume prediction can be formulated as a pure Time Series Forecasting (TSF) problem, in which the temporal evolution of a data rate in the near future is forecast. In contrast to that, regression algorithms use correlated statistical traffic features to approximate a mapping between the features and the expected traffic volume. There are other, related problems such as the prediction or classification of the lifetime of a flow or its size. Please refer to [1] for an overview of existing literature on the topic.
The task for the student is to design and implement a traffic prediction algorithm based on Machine Learning (ML), which includes
It is possible to focus on specific aspects of the task.
Supervisor: Johannes Busch
With emerging Vehicle-to-Infrastructure (V2I) communication technologies, such as IEEE 802.11p or 5G, arises the opportunity for an intelligent traffic infrastructure that is able to quickly adapt to the current state of the traffic system.Through the exchange of information between individual vehicles and the traffic infrastructure, traffic signaling could be tailored to individual vehicles. Furthermore the traffic infrastructure could advise drivers on appropriate driving velocities. These measures could improve traffic flow and therefore mitigate environmental and economic repercussions of traffic systems. However, the optimization of large distributed systems under the availability of intricate knowledge about the system state is not trivial and novel control paradigms need to be explored. A promising candidate could be Deep Reinforcement Learning (DRL), which has shown to be able to solve many intricate control problems by learning from interaction and has also successfully been applied to traffic control systems.
In this Diploma/Master thesis, a joint optimization of traffic signaling and speed advisory with DRL should be explored. Subtasks include:
Prospective students should have previous knowledge in the Python programming language and a keen interest and previous experience in Machine Learning (preferably DRL).
Network congestion [2] occurs when a node or link has to process more traffic than it can handle, causing undesired delays and congestion losses in queues. Congestion Control (CC) algorithms, either running in endpoints or within the network, mitigate congestion and are essential for sustaining the quality of service of networks today.
In recent years, CC algorithms made use of Machine Learning (ML) to improve their control decisions. Interested students can work on an ML-aided (supervised learning) sender-side congestion control algorithm.
Supervisor: Christian Vielhaus and Johannes Busch
In 5G, distributed core networks that are orchestrated with Software-Defined Networking (SDN) provide Computing In the Network (COIN) at edge clouds. SDN enables network operators to adapt these networks in real time and a key challenge is to route flows dynamically to improve network performance (e.g. load balancing, reduce propagation delay after a handover, …) and fulfill service requirements (e.g. network function required along the path).
ML can help or control the decision when and how routes have to change by either making predictions that are used by routing algorithms or by directly updating forwarding tables with Reinforcement Learning (RL).
[1] A Comprehensive Survey on Machine Learning for Networking: Evolution, Applications and Research Opportunities [2] Network congestion