LP-LM is a novel system that grounds language model responses in verifiable facts using knowledge bases and logic programming, eliminating hallucinations.

• Logic-based verification ensures all generated answers are reliable and factually accurate
• Semantic parsing with Prolog creates structured understanding of natural language questions
• Knowledge base integration grounds responses in established facts rather than probabilistic generation
• Enhanced security prevents misinformation and ensures trustworthy AI communications

This research addresses critical security concerns by preventing AI systems from generating false information that could lead to harmful decision-making in sensitive domains.

LP-LM: No Hallucinations in Question Answering with Logic Programming