MC-RAG System: A Structure-Driven RAG System for Multi-Constraint Queries

2026-07-11

AI Summary

This work addresses the limitations of conventional retrieval-augmented generation (RAG) systems, which often violate constraints, produce factually inconsistent outputs, or generate hallucinations when handling complex multi-constraint queries. To overcome these issues, the authors propose a novel structure-driven RAG paradigm that formulates multi-constraint queries as subgraph matching tasks over knowledge graphs. By integrating semantic and structural embeddings and introducing path-level indexing for efficient retrieval, the approach enables end-to-end interactive query processing in domains such as healthcare and encyclopedic knowledge. The method not only supports transparent visualization of constraint parsing and structural matching but also significantly enhances the accuracy, constraint adherence, and interpretability of generated responses.

Abstract

Retrieval-Augmented Generation (RAG) systems are widely adopted in question answering, yet they often fail to satisfy complex multi-constraint queries, leading to constraint violations, factual inconsistencies, or hallucinations. We present Structure-Driven RAG System for Multi-Constraint Queries(MC-RAG), a structure-driven RAG system that reformulates retrieval as a subgraph matching problem over a knowledge graph. By integrating semantic and structural embeddings with path-level indexing, MC-RAG performs interpretable, structure-aware, and constraint-consistent retrieval and generation. During the demonstration, participants can input medical or encyclopedic multi-constraint queries, visualize how the system parses constraints, performs structural matching, and generates answers, thereby experiencing an end-to-end, interactive, and explainable RAG pipeline. A demo video is available at https://youtu.be/J8kahzmAnu0.

Problem

• Retrieval-Augmented Generation
• multi-constraint queries
• constraint violation
• factual inconsistency
• hallucination

Innovation

• Structure-Driven RAG
• Subgraph Matching
• Multi-Constraint Queries
• Knowledge Graph
• Constraint-Consistent Generation