# Discovering Relationships in Data Lakes Using Large Language Models: An Industrial Case

arXiv: https://arxiv.org/abs/2608.26750 · HTML: https://mlsanity.com/paper/2608.26750/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 12 of its 30 references that are in the map.
- First posted on arXiv 27 Aug 2026.

## Builds on (most cited first, 12 of 12)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers](https://mlsanity.com/paper/2002.10957/) (2020)
- [MPNet: Masked and Permuted Pre-training for Language Understanding](https://mlsanity.com/paper/2004.09297/) (2020)
- [MTEB: Massive Text Embedding Benchmark](https://mlsanity.com/paper/2210.07316/) (2022)
- [Valentine: Evaluating Matching Techniques for Dataset Discovery](https://mlsanity.com/paper/2010.07386/) (2020)
- [Dataset Discovery in Data Lakes](https://mlsanity.com/paper/2011.10427/) (2020)
- [Schema Matching with Large Language Models: an Experimental Study](https://mlsanity.com/paper/2407.11852/) (2024)
- [Magneto: Combining Small and Large Language Models for Schema Matching](https://mlsanity.com/paper/2412.08194/) (2024)
- [TARGET: Benchmarking Table Retrieval for Generative Tasks](https://mlsanity.com/paper/2505.11545/) (2025)
- [ReMatch: Retrieval Enhanced Schema Matching with LLMs](https://mlsanity.com/paper/2403.01567/) (2024)
- [DeepJoin: Joinable Table Discovery with Pre-trained Language Models](https://mlsanity.com/paper/2212.07588/) (2022)
- [Evaluating Joinable Column Discovery Approaches for Context-Aware Search](https://mlsanity.com/paper/2510.24599/) (2025)

## Abstract (arXiv)

Data lakes rely on metadata to remain usable, yet this meta data is often limited or weakly informative for column relationship discovery, especially in ERP-derived datasets with coded or abbreviated schema labels. We propose ColRel, a two-stage method that builds column embeddings from metadata and data available at ingestion time. In difficult cases, such as coded schemata, business dictionaries help better interpret column names and support the generation of short natural-language descriptions used in the second stage. Experiments on public benchmarks and an industrial ERP dataset show that ColRel is particularly effective in semantically related, weak-signal settings.