Abstract
The evolution toward sixth-generation (6G) networks introduces agentic communication, in which autonomous software agents interact and execute actions based on real-time environmental data. Large language models (LLMs) provide reasoning capability for such agents, but deploying an individual LLM for each agent can be operationally and economically infeasible. This disclosure presents a consolidated, context-aware LLM architecture that centralizes LLM resources within a 6G core and logically separates the resources into control plane (CP) and user plane (UP) instances. Using a shared resource model and a context-tagging mechanism based on device type, radio access technology (RAT), and function type, the architecture supports high-performance autonomous network operations while reducing hardware footprint and total cost of ownership (TCO).
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Barai, Swapnil; Shekhar, Ravi; Parikh, Dishant; and Rungta, Gaurav Kumar, "OPTIMIZED, CONTEXT-AWARE LARGE LANGUAGE MODEL (LLM) ARCHITECTURE FOR ARTIFICIAL INTELLIGENCE (AI)-NATIVE SIXTH-GENERATION (6G) NETWORKS", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10232