Communication-Efficient Decentralized Intelligence in 6G Systems

The evolution of 6G will be shaped by the convergence of advanced communication technologies and artificial intelligence (AI)-driven systems.

A growing portion of network traffic will consist of data supporting AI and machine learning applications, generated by sources such as autonomous vehicles, drone swarms, camera networks, augmented and virtual reality, and foundation models.

This traffic differs fundamentally from the audio, video, and web traffic that has dominated previous wireless generations.

These differences are evident not only in the velocity, volume, and traffic patterns of data streams but also in their sensitivity to delays and losses.

Massive machine-learning tasks, such as decentralized learning, edge computing, and real-time decision-making, will require innovations at both the communication and AI levels to operate efficiently at scale.To address these challenges, we propose a research environment that brings together three of Sweden´s leading research groups in information theory, machine learning, and wireless communications, with a strong track record in both fundamental research and technology innovations.

Our aim is to explore the fundamental limits of distributed learning and decision-making, develop efficient communication strategies for 6G networks, and design machine-learning algorithms capable of handling the network imperfections that will remain impractical to resolve at the physical layer also in next generation networks.