Structured text improves table reading and question answering tasks
Tables Decoded: DELTA for Structure, TARQA for Understanding
Computer Vision and Pattern RecognitionMachine Learning
Summary
Understanding tables in documents is important but tricky, especially when tables are in images and different languages. The authors developed DELTA, a way to turn tables into a clear written format that captures both layout and content accurately. They also created TARQA, a tool that uses large language models to answer questions about these tables using that format. Their new approach works well across several datasets and languages, including Hindi, and performs better than many previous methods.
What this means in practice
- •For document processing teams: Extract accurate and language-independent table data from documents for further automated analysis using structured textual representations.
- •For customer support platforms: Improve automated answering of questions about tabular data in documents by using fine-tuned language models on structured table formats.
Authors
Jahanvi Rajput, Dhruv Kudale, Saikiran Kasturi, Utkarsh Verma, Ganesh Ramakrishnan
Abstract
Table understanding is a core task in document intelligence, encompassing two key subtasks: table reconstruction and table visual question answering (TabVQA). While recent approaches predominantly rely on vision- language models (VLMs) operating on table images, we propose a more scalable and effective alternative based on structured textual representations. These representations are easier to process, align more naturally with LLMs, and eliminate the need for language-specific visual encoders, making them particularly suitable for multilingual documents. We present DELTA, which separates physical structure recognition, logical structure recognition, and OCR to extract both layout and content accurately. DELTA outputs tables in Optimised Table Structure Language (OTSL), a compact and unified format that encodes cell arrangements and textual content. On table structure recognition (TSR), DELTA achieves TEDS- Structure scores comparable with state-of-the-art methods across FinTabNet, PubTabNet, and PubTables-1M. We further establish its robustness on non-English tables through our curated Hindi benchmark, TORQUE. Building on this, we introduce TARQA, an LLM fine-tuned on OTSL sequences. Our approach yields gains of 9.3 p.p. on WTQ (TabQA) and 9.2 p.p. on FinTabNetQA (TabVQA), respectively. On TORQUE, our method ranks second among all VLMs and DELTA + LLM variants. We release our code, models, and benchmark at: https://github.com/Tihiitborg/Tables-Decoded