This paper introduces Agentic-V2X, an architecture designed to use small language models for deadline-aware vehicle-to-everything (V2X) scheduling in 5G/6G networks. The system employs a small, local language model to generate policies, which are then validated and executed by a lightweight controller. Evaluations using ns-3/ns3-ai demonstrate that Agentic-V2X can produce valid policies and maintain competitive performance, particularly in critical reliability scenarios, though it does not consistently outperform the strongest static policies. AI
IMPACT Proposes a novel architecture for integrating small LLMs into real-time network scheduling, addressing latency and control concerns.
RANK_REASON Academic paper detailing a novel architecture for LLM agents in network scheduling. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
