PolyUQuest: Verifiable Structure-Aware Web RAG over Heterogeneous Graphs

2026-07-09 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Advanced, quick

Summary

PolyUQuest is a novel, verifiable, structure-aware web Retrieval-Augmented Generation (RAG) framework designed to overcome the limitations of existing systems that treat web pages as flat text. It constructs a heterogeneous graph integrating hyperlink topology, DOM hierarchy, and entity-relation knowledge across web pages. A two-tier router intelligently dispatches queries to one of three retrieval modes: direct block retrieval, cross-page graph traversal, or multi-hop entity reasoning, based on the query's structural needs. A key feature is its full verifiability, where every cited answer block provides its source page, heading path, and entity links, enabling users to trace claims to their structural evidence. Evaluated on the Hong Kong Polytechnic University's official websites, comprising 4,240 pages, 31,086 DOM blocks, 29,119 entities, and 37,680 relations, PolyUQuest significantly outperforms current RAG systems in answer correctness, coverage, and faithfulness, while also reducing LLM token consumption. It is slated for deployment as a student-facing QA service at PolyU.

Key takeaway

For AI Engineers developing RAG systems over complex, structured web content, PolyUQuest demonstrates that treating web pages as flat text severely limits performance and verifiability. You should explore integrating heterogeneous graphs that unify DOM hierarchy, hyperlink topology, and entity-relation knowledge to significantly improve answer correctness, coverage, and faithfulness. This approach also reduces LLM token consumption and provides critical verifiability, allowing users to trace claims directly to their structural evidence.

Key insights

PolyUQuest enhances RAG by using heterogeneous graphs for structure-aware, verifiable web information retrieval.

Principles

• Integrate web structure (DOM, hyperlinks) into RAG.
• Enable answer verifiability via structural evidence.
• Route queries based on structural complexity.

Method

Construct a heterogeneous graph from web pages, unifying hyperlink topology, DOM hierarchy, and entity-relation knowledge, then use a two-tier router to dispatch queries to specific retrieval modes.

In practice

• Implement graph-based RAG for complex web data.
• Provide source page, heading path, and entity links for verifiability.
• Optimize LLM token use with structure-aware retrieval.

Topics

• Retrieval-Augmented Generation
• Heterogeneous Graphs
• Web Data Structure
• Verifiable AI
• DOM Hierarchy
• LLM Efficiency

Best for: Research Scientist, AI Architect, NLP Engineer, AI Scientist, Machine Learning Engineer, AI Engineer

Related on AIssential

• PolyUQuest: Verifiable Structure-Aware Web RAG over Heterogeneous Graphs — PolyUQuest unifies web structure into a heterogeneous graph for verifiable, efficient, and accurate RAG.
• Architectural patterns for graph-enhanced RAG: Moving beyond vector search in production — Graph-enhanced RAG improves LLM accuracy by integrating structural data from graph databases with semantic vector search.
• FAIR GraphRAG: A Retrieval-Augmented Generation Approach for Semantic Data Analysis — Integrating FAIR Digital Objects into graph-based RAG significantly enhances data findability, interoperability, and question-answering performance.
• Don’t Put Everything into RAG: Building a Safe Candidate Layer for Web Sources — Directly feeding unverified internet data into RAG systems introduces significant risks and should be avoided.
• Reducing Hallucinations in Complex Question Answering using Simple Graph-based Retrieval-Augmented Generation (long version) — Using a lightweight graph structure in RAG systems significantly reduces LLM hallucinations and improves factual correctness.
• How Data Graphs Built a True Hybrid Graph RAG Platform — True hybrid RAG combines structured graph data with semantic vector search, dynamically orchestrated by an AI agent for superior results.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.