|Summary||Future mobile communication networks are characterized by unprecedented
levels of complexity, which exceed the scope of current (mostly
model-based) approaches to network management and operations.
Data-driven approaches, learning algorithms and emerging artificial
intelligence concepts provide powerful alternative techniques,
but their use in mobile networks raises formidable challenges,
for example in terms of the massive amounts of data needed to ensure
satisfactory performance and possibly slow convergence.
The joint use of data-driven and model-based approaches
(e.g. “expert-knowledge-aided deep learning”) provides a promising
paradigm to overcome these challenges.
The goal of this PhD project is to explore the potential benefits
of combining data-driven and model-based approaches,
provide benchmarks of achievable gains in various scenarios,
and develop concrete algorithms for specific use-cases in the context
of Beyond 5G mobile networks.
|Supervisors||Sem Borst (TU/e)|
Eindhoven University of Technology (TU/e)