This proposal provides a clean-slate approach to the problem of context-aware resource management in 5G heterogeneous networks.
This is achieved by exploring a dimension that has been often overlooked so far -- users' context, via a novel, self-organizing
framework that allows to: (i)- profile users dynamically using context information learned from dimensions such as smartphone,
geographical and social information, (ii)- exploit the extracted profiles to develop optimized resource management techniques over
time, frequency, and space. This leads to a holistic framework that continuously adapts the network operation to users' context,
yielding efficient, context-aware and self-organizing networks. To date, the classical approach to tackle these problems is via Monte-
Carlo simulations, offline processing or heuristics.