TabRAG: Tabular Document Retrieval via Structured Language Representations

Jacob Si, Mike Qu, Michelle Lee, Yingzhen Li

Submission Type: Short paper (4 pages)

Keywords: RAG, Vision Language Model, Large Language Model

Abstract: Ingesting data for Retrieval-Augmented Generation (RAG) involves either fine-tuning the embedding model directly on the target corpus or parsing documents for embedding model encoding. The former, while accurate, incurs high computational hardware requirements, while the latter suffers from suboptimal performance when extracting tabular data. In this work, we address the latter by presenting TabRAG, a parsing-based RAG pipeline designed to tackle table-heavy documents via structured language representations. TabRAG achieves state-of-the-art performance over existing parsing-based methods for generation and retrieval.

Submission Number: 22