Knowledge-Guided LLM Framework for 6G Network Service Orchestration

This framework integrates grammar-constrained semantic parsing with a knowledge-graph clarification loop to convert ambiguous natural language into valid 6G service profiles. In testing, it achieved 89% classification accuracy and 95% policy validity, significantly outperforming pure LLM baselines in contradiction detection.