# Pneuma: Leveraging LLMs for Tabular Data Representation and Retrieval in an End-to-End System

arXiv: https://arxiv.org/abs/2504.09207 · HTML: https://mlsanity.com/paper/2504.09207/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- Placed in RAG robustness and adaptive retrieval, one of 15 directions in Retrieval-augmented generation.
- Builds on 15 of its 20 references that are in the map.
- First posted on arXiv 12 Apr 2025.

## Placed in

- Retrieval-augmented generation / [RAG robustness and adaptive retrieval](https://mlsanity.com/rag/rag-robustness-and-adaptive-retrieval/)

## Builds on (most cited first, 15 of 15)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [Datasheets for Datasets](https://mlsanity.com/paper/1803.09010/) (2018)
- [MTEB: Massive Text Embedding Benchmark](https://mlsanity.com/paper/2210.07316/) (2022)
- [TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data](https://mlsanity.com/paper/2005.08314/) (2020)
- [FeTaQA: Free-form Table Question Answering](https://mlsanity.com/paper/2104.00369/) (2021)
- [ReAcTable: Enhancing ReAct for Table Question Answering](https://mlsanity.com/paper/2310.00815/) (2023)
- [Open Domain Question Answering over Tables via Dense Retrieval](https://mlsanity.com/paper/2103.12011/) (2021)
- [An Analysis of Fusion Functions for Hybrid Retrieval](https://mlsanity.com/paper/2210.11934/) (2022)
- [Retrieving Complex Tables with Multi-Granular Graph Representation Learning](https://mlsanity.com/paper/2105.01736/) (2021)
- [Ad Hoc Table Retrieval using Semantic Similarity](https://mlsanity.com/paper/1802.06159/) (2018)
- [Web Table Extraction, Retrieval and Augmentation: A Survey](https://mlsanity.com/paper/2002.00207/) (2020)
- [Solo: Data Discovery Using Natural Language Questions Via A Self-Supervised Approach](https://mlsanity.com/paper/2301.03560/) (2023)
- [Data Market Platforms: Trading Data Assets to Solve Data Problems](https://mlsanity.com/paper/2002.01047/) (2020)
- [CMDBench: A Benchmark for Coarse-to-fine Multimodal Data Discovery in Compound AI Systems](https://mlsanity.com/paper/2406.00583/) (2024)

## Built on by (most cited first, 23 of 23)

- [A Survey of Data Agents: Emerging Paradigm or Overstated Hype?](https://mlsanity.com/paper/2510.23587/) (2025)
- [Can LLMs Clean Up Your Mess? A Survey of Application-Ready Data Preparation with LLMs](https://mlsanity.com/paper/2601.17058/) (2026)
- [AGRAG: Advanced Graph-based Retrieval-Augmented Generation for LLMs](https://mlsanity.com/paper/2511.05549/) (2025)
- [Arming Data Agents with Tribal Knowledge](https://mlsanity.com/paper/2602.13521/) (2026)
- [Towards Universal Tabular Embeddings: A Benchmark Across Data Tasks](https://mlsanity.com/paper/2604.21696/) (2026)
- [A Survey of LLM × DATA](https://mlsanity.com/paper/2505.18458/) (2025)
- [QCFuse: Query-Aware Cache Fusion via Compressed View for Efficient RAG Serving](https://mlsanity.com/paper/2606.05875/) (2026)
- [Topo-RAG: Topology-aware retrieval for hybrid text-table documents](https://mlsanity.com/paper/2601.10215/) (2026)
- [GraphSeek: Next-Generation Graph Analytics with LLMs](https://mlsanity.com/paper/2602.11052/) (2026)
- [Fine-Grained Table Retrieval Through the Lens of Complex Queries](https://mlsanity.com/paper/2603.07146/) (2026)
- [Blue Data Intelligence Layer: Streaming Data and Agents for Multi-source Multi-modal Data-Centric Applications](https://mlsanity.com/paper/2604.15233/) (2026)
- [Innovative tokenisation of structured data for LLM training](https://mlsanity.com/paper/2508.01685/) (2025)
- [Cardinality Estimation for High Dimensional Similarity Queries with Adaptive Bucket Probing](https://mlsanity.com/paper/2604.04603/) (2026)
- [Demonstration of Pneuma-Seeker: Agentic System for Reifying and Fulfilling Information Needs on Tabular Data](https://mlsanity.com/paper/2604.14422/) (2026)
- [RedParrot: Accelerating NL-to-DSL for Business Analytics via Query Semantic Caching](https://mlsanity.com/paper/2604.22758/) (2026)
- [Retrieve Only Relevant Tables Whether Few or Many: Adaptive Table Retrieval Method](https://mlsanity.com/paper/2605.18766/) (2026)
- [Sophrosyne: Agentic Exploration of Relational Data Systems Needs Moderation](https://mlsanity.com/paper/2605.30862/) (2026)
- [SANA: What Matters for QA Agents over Massive Data Lakes?](https://mlsanity.com/paper/2606.13904/) (2026)
- [Stellar: Scalable Multimodal Document Retrieval for Natural Language Queries](https://mlsanity.com/paper/2606.19960/) (2026)
- [Semantically Similar, Logically Distinct: Diagnosing the Semantic-Answerability Gap in Table RAG](https://mlsanity.com/paper/2607.17742/) (2026)
- [Polaris: Learning to Generate Table Descriptions from Retrieval Feedback](https://mlsanity.com/paper/2608.17171/) (2026)
- [Open Tabular Insight Extraction: Where Do We Stand, and Where Should We Go?](https://mlsanity.com/paper/2609.07629/) (2026)
- [Discovery-Driven Integration of Disjoint Tables via Text](https://mlsanity.com/paper/2609.26658/) (2026)

## Abstract (arXiv)

Finding relevant tables among databases, lakes, and repositories is the first step in extracting value from data. Such a task remains difficult because assessing whether a table is relevant to a problem does not always depend only on its content but also on the context, which is usually tribal knowledge known to the individual or team. While tools like data catalogs and academic data discovery systems target this problem, they rely on keyword search or more complex interfaces, limiting non-technical users' ability to find relevant data. The advent of large language models (LLMs) offers a unique opportunity for users to ask questions directly in natural language, making dataset discovery more intuitive, accessible, and efficient. In this paper, we introduce Pneuma, a retrieval-augmented generation (RAG) system designed to efficiently and effectively discover tabular data. Pneuma leverages large language models (LLMs) for both table representation and table retrieval. For table representation, Pneuma preserves schema and row-level information to ensure comprehensive data understanding. For table retrieval, Pneuma augments LLMs with traditional information retrieval techniques, such as full-text and vector search, harnessing the strengths of both to improve retrieval performance. To evaluate Pneuma, we generate comprehensive benchmarks that simulate table discovery workload on six real-world datasets including enterprise data, scientific databases, warehousing data, and open data. Our results demonstrate that Pneuma outperforms widely used table search systems (such as full-text search and state-of-the-art RAG systems) in accuracy and resource efficiency.