MemQ: Enhancing Knowledge Graph Question Answering with Memory-Augmented Query Reconstruction

MemQ: Enhancing Knowledge Graph Question Answering with Memory-Augmented Query Reconstruction

LLMs have shown strong performance in Knowledge Graph Question Answering (KGQA) by leveraging planning and interactive strategies to query knowledge graphs. Many existing approaches rely on SPARQL-based tools to retrieve information, allowing models to generate accurate answers. Some methods enhance LLMs’ reasoning abilities by constructing tool-based reasoning paths, while others employ decision-making frameworks that use environmental feedback…