From Chunks to Functional Evidence: Function-Aware Retrieval for EDA Documentation QA

2026-10-06

AI Summary

This study addresses the issue of answer deviation in Electronic Design Automation (EDA) document retrieval caused by the misalignment between queries and knowledge organization. To this end, we propose a function-aware evidence organization method based on hypergraphs. This approach reconstructs Retrieval-Augmented Generation (RAG) retrieval units by aggregating heterogeneous artifacts into functional units, replacing conventional isolated chunking and pairwise relation modeling with hyperedge-based representation. Furthermore, it achieves precise evidence selection through query-unit alignment encoder training combined with a hybrid retrieval and reranking mechanism. Experimental results demonstrate substantial improvements, achieving a 37.1% increase in ROUGE-L on the EDADocEval-QA dataset and a 30.0% gain on the ORD-MMBench benchmark.

Abstract

Retrieval-Augmented Generation (RAG) is widely used to ground answers in documents. For complex technical documentation, however, the primary bottleneck is often not model reasoning but a mismatch between a query and the way knowledge is organized for retrieval. This mismatch is pronounced in Electronic Design Automation (EDA) documentation, where the information needed for an answer is scattered across heterogeneous yet tightly coupled artifacts. We therefore redesign the basic retrieval unit of RAG. Instead of operating on isolated chunks or binary relations, we collect typed artifacts into EDA functional units. Each unit is recorded as a hyperedge with links to its source chunks. We then train an encoder to align queries with functional units and combine unit retrieval with direct chunk retrieval. After mapping the selected units back to their sources, a unified reranker chooses the evidence given to the generator. On the newly constructed EDADocEval-QA dataset, our method improves ROUGE-L by 37.1% over Chunk RAG and 55.6% over the strongest graph baseline. On the public ORD-MMBench benchmark, it improves ROUGE-L by 30.0% over the strongest baseline. These results support function-aware evidence organization in the evaluated EDA documentation settings.

Problem

• Retrieval-Augmented Generation
• Electronic Design Automation
• documentation QA
• information retrieval
• knowledge organization

Innovation

• Retrieval-Augmented Generation
• Function-Aware Retrieval
• Hyperedge
• Electronic Design Automation
• Unified Reranker