analiza aiqbench
HiQA addresses a critical bottleneck in Retrieval-Augmented Generation by improving accuracy when knowledge bases contain numerous semantically similar documents through hierarchical metadata augmentation. The proposed multi-route retriever effectively combines semantic, lexical, and structural signals to enhance answer quality in complex, domain-specific scenarios. This work provides a practical framework for optimizing RAG systems where traditional retrieval methods struggle with document redundancy.
HiQA: A Hierarchical Contextual Augmentation RAG for Multi-Documents QA
arXiv:2402.01767v4 Announce Type: replace-cross Abstract: Retrieval-Augmented Generation (RAG) significantly improves document-based question answering by integrating external documents during generation. However, retrieval accuracy can degrade when the knowledge base contains many semantically and structurally similar documents. We introduce HiQA, a practical hierarchical contextual augmentation framework for multi-document question answering (MDQA). HiQA enriches text chunks with cascading document metadata, such as document titles and section paths, so that retrieval can use both local content and document structure. The framework also uses a multi-route retriever that combines semantic, lexical, and keyword/entity signals. We further introduce MasQA, a benchmark designed to evaluate MDQA systems in realistic similar-document settings. Experiments show that HiQA improves retrieval and answer quality on MasQA and remains competitive on public MDQA benchmarks, while its benefits are strongest for structured, domain-specific, highly similar document collections.
Abstract:Retrieval-Augmented Generation (RAG) significantly improves document-based question answering by integrating external documents during generation. However, retrieval accuracy can degrade when the knowledge base contains many semantically and structurally similar documents. We introduce HiQA, a practical hierarchical contextual augmentation framework for multi-document question answering (MDQA). HiQA enriches text chunks with cascading document metadata, such as document titles and section paths, so that retrieval can use both local content and document structure. The framework also uses a multi-route retriever that combines semantic, lexical, and keyword/entity signals. We further introduce MasQA, a benchmark designed to evaluate MDQA systems in realistic similar-document settings. Experiments show that HiQA improves retrieval and answer quality on MasQA and remains competitive on public MDQA benchmarks, while its benefits are strongest for structured, domain-specific, highly similar document collections.
Metadata correction: the arXiv author metadata has been updated to include Xinjian Chen, consistent with the author list already present in the v3 manuscript files. The manuscript content is unchanged
Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2402.01767 [cs.CL]
(or arXiv:2402.01767v4 [cs.CL] for this version)
Text integral sub licența deschisă a publicației · 177 cuvinte
Subiecte conexe
A Unified Framework for Context-Aware and Relation-Aware Graph Retrieval-Augmented Generation
arXiv · 17 iun. 2026
Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers
Hugging Face · 18 aug. 2026
RH-RAG: Trustworthy Long-Form Generation for Privacy-Constrained Settings
arXiv · 4 aug. 2026
From code to diagrams: Agentic architecture documentation with Amazon Bedrock AgentCore
AWS · 2 sept. 2026