Bayesian trust networks, source-aware knowledge distillation, and multi-agent debate protocols can estimate and reconcile confidence scores across modalities and sources to detect and resolve conflicts in knowledge-graph-enhanced Large Language Models.

Authors

• Person: Not available Organization: arXiv
  Large Language Models Meet Knowledge Graphs for Question ...

Sources

• Large Language Models Meet Knowledge Graphs for Question ... arxiv.org arXiv via serper

Referenced by nodes (1)

• Large Language Models concept