AI and foundation models pave the way for terahertz ultra-massive multiple-input multiple-output in 6G

https://www.eurekalert.org/news-releases/1122510

"6G need for >1,024 antennas creates computational hurdles, near-/ far-field effects traditional formulas don't capture well, difficult to physically get accurate Channel State Information when dealing with many arrays of subarrays... overcome using AI/ foundation models shifting from rigid to more flexible AI-driven formulas: model-driven deep learning, 1 physical layer foundation model/ many tasks, LLMs for network: management/ optimization searches/ directing other parts system to manage traffic flow autonomously"

Related:

Task-driven AI-native architecture for 6G networks
https://www.eurekalert.org/news-releases/1123996

6G SAGINs designed for unmanned operations with SC³ loop as core unit
https://www.eurekalert.org/news-releases/1123974

Generative AI paves new way for semantic communication in 6G networks
https://www.eurekalert.org/news-releases/1123972

Cooperative sensing paves the way for 6G ISAC development
https://www.eurekalert.org/news-releases/1123995

Leveraging data correlation extends robot swarm lifetime in 6G edge computing
https://www.eurekalert.org/news-releases/1123998

WEIT paves new way for 6G environment intelligence communication
https://www.eurekalert.org/news-releases/1123988