International Journal of Science, Engineering and TechnologyAuthors: Assistant Professor Ms. K.Nivetha, Uthraa Ujjwal pu
Abstract: This project introduces a chatbot solution that combines Retrieval-Augmented Generation (RAG) with a Knowledge Graph (KG) to respond to inquiries related to Indian law. The system intelligently adjusts to the complexity of user questions to maximize both efficiency and accuracy. For direct and simple queries, the RAG module fetches pertinent legal documents and delivers concise responses. In contrast, for intricate and layered questions, the system utilizes a multi-hop reasoning technique through the knowledge graph to produce precise and context-rich answers. To broaden reach and inclusiveness, the chatbot is equipped with multilingual capabilities, allowing individuals from various linguistic backgrounds across India to engage in their native language. This blended architecture promotes computational efficiency by activating the knowledge graph module only when required, thereby conserving system resources while maintaining high response quality. The chatbot ensures smooth and intuitive access to legal information—offering fast replies for basic questions and thorough explanations for more complex legal matters. By harnessing the complementary advantages of RAG and KG, this system seeks to transform legal support services by enhancing availability, reliability, and user-friendliness for both legal experts and the general.
DOI: https://doi.org/10.5281/zenodo.15568853
