ESSENCE – Efficient Self-Supervised Machine Learning for Adaptive Wireless Communication Systems
This research project explores how self-supervised learning can make future wireless communication systems more adaptable, efficient, robust, and scalable.
About ESSENCE
Today’s wireless networks already use machine learning for specific tasks. However, recent progress in other fields shows that more general and flexible learning models can adapt better to new situations. ESSENCE builds on this idea by studying how such models can be trained directly from data generated by wireless systems, without relying on large amounts of labeled data.
The project focuses on learning from channel state information, which describes how radio signals behave in different environments. By using this information, ESSENCE aims to develop compact learning models that can adapt locally to changing conditions and support key functions in wireless networks.
Overall, the project lays the groundwork for intelligent wireless systems that can learn continuously and cope with the increasing complexity of future networks.
