IA-RAG: Interval-Algebra-Driven Temporal Reasoning for Dynamic Knowledge Retrieval

Retrieval-Augmented Generation (RAG) has shown strong effectiveness in grounding Large Language Models (LLMs) with external knowledge. However, existing RAG and Graph RAG frameworks largely treat knowledge as static or associate time with coarse-grained timestamps or metadata, failing to capture ric...