Query Symbolically or Retrieve Semantically? A Dataset and Method for Semi-Structured Question Answering
The article discusses a new framework called DualGraph designed for semi-structured question answering. It combines semantic retrieval and symbolic querying to improve the effectiveness of retrieval-augmented generation systems. The authors also introduce a benchmark dataset, SpecsQA, to evaluate the performance of their method against existing approaches.
- ▪DualGraph represents documents through a Textual Knowledge Graph for semantic retrieval and a Symbolic Knowledge Graph for symbolic querying.
- ▪The framework aims to address the limitations of current retrieval methods on semi-structured corpora.
- ▪Experiments demonstrate that DualGraph outperforms state-of-the-art methods in various question-answering scenarios.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.27164 (cs) [Submitted on 26 May 2026] Title:Query Symbolically or Retrieve Semantically? A Dataset and Method for Semi-Structured Question Answering Authors:Mateusz Czyżnikiewicz, Ryszard Tuora, Adam Kozakiewicz, Tomasz Ziętkiewicz, Mateusz Galiński, Michał Godziszewski, Michał Karpowicz, Timothy Hospedales, Cristina Cornelio View a PDF of the paper titled Query Symbolically or Retrieve Semantically? A Dataset and Method for Semi-Structured Question Answering, by Mateusz Czy\.znikiewicz and 8 other authors View PDF HTML (experimental) Abstract:Retrieval-Augmented Generation (RAG) systems for question answering typically retrieve evidence by semantic similarity between the query and document chunks.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
