6G Goal-Oriented AI-enabled Learning and Semantic Communication Networks

Wireless mobile communication has continually evolved towards higher data rates, with 5G expanding its scope to include massive and ultra-reliable low-latency links.

The underlying quest in the wireless evolution has been to solve the technical problem of reliable data exchange between two end-points. 6G-GOALS will take the wireless system design to its next stage by considering the significance, relevance, and value of the transmitted data, transforming the potential of the emerging AI/ML-based architectures into a semantic and goal-oriented communication paradigm.

Two trends corroborate the timeliness of 6G-GOALS: (1) the burden on wireless networks by data flows with low semantic content or relevance for the end goal; (2) the increased AI capability of network nodes and devices to extract ‘meaning’ and intention from unreliable data flows.

These trends underpin the two main objectives of 6G-GOALS: (1) to reduce data traffic by conveying only the most relevant information; (2) to design data-efficient, robust, and resilient protocols that can adapt to network conditions and communication objectives using modern AI/ML techniques.

The research breakthroughs and innovation of 6G-GOALS are three-fold: First, 6G-GOALS will develop AI/ML-empowered semantic data representation, sensing, and compression algorithms combining data-and-model-driven approaches, and work towards exploiting untapped gains from AI-based joint source-channel coding.

Second, 6G-GOALS will introduce semantic-oriented solutions for supporting distributed reasoning and time-sensitive communication, generalizing low- latency of 5G by tailoring communication to the actual goal.

Finally, 6G-GOALS will introduce wireless technologies for sustainability in energy efficiency, EMF exposure, and spectrum management, by defining the concept of semantic cognitive radio.