# Discovery-Driven Integration of Disjoint Tables via Text

arXiv: https://arxiv.org/abs/2609.26658 · HTML: https://mlsanity.com/paper/2609.26658/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 34 references that are in the map.
- First posted on arXiv 22 Sep 2026.

## Builds on (most cited first, 17 of 17)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data](https://mlsanity.com/paper/2005.08314/) (2020)
- [Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free](https://mlsanity.com/paper/2505.06708/) (2025)
- [LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics](https://mlsanity.com/paper/2511.08544/) (2025)
- [Data Programming: Creating Large Training Sets, Quickly](https://mlsanity.com/paper/1605.07723/) (2016)
- [FEVEROUS: Fact Extraction and VERification Over Unstructured and Structured information](https://mlsanity.com/paper/2106.05707/) (2021)
- [SANTOS: Relationship-based Semantic Table Union Search](https://mlsanity.com/paper/2209.13589/) (2022)
- [ProTrix: Building Models for Planning and Reasoning over Tables with Sentence Context](https://mlsanity.com/paper/2403.02177/) (2024)
- [Pneuma: Leveraging LLMs for Tabular Data Representation and Retrieval in an End-to-End System](https://mlsanity.com/paper/2504.09207/) (2025)
- [Mixed-modality Representation Learning and Pre-training for Joint Table-and-Text Retrieval in OpenQA](https://mlsanity.com/paper/2210.05197/) (2022)
- [Multi-modal Retrieval of Tables and Texts Using Tri-encoder Models](https://mlsanity.com/paper/2108.04049/) (2021)
- [Termite: A System for Tunneling Through Heterogeneous Data](https://mlsanity.com/paper/1903.05008/) (2019)
- [Graph integration of structured, semistructured and unstructured data for data journalism](https://mlsanity.com/paper/2007.12488/) (2020)
- [Cross Modal Data Discovery over Structured and Unstructured Data Lakes](https://mlsanity.com/paper/2306.00932/) (2023)
- [Ver: View Discovery in the Wild](https://mlsanity.com/paper/2106.01543/) (2021)
- [Snoopy: Effective and Efficient Semantic Join Discovery via Proxy Columns](https://mlsanity.com/paper/2502.16813/) (2025)
- [KGLiDS: A Platform for Semantic Abstraction, Linking, and Automation of Data Science](https://mlsanity.com/paper/2303.02204/) (2023)

## Abstract (arXiv)

Integrating heterogeneous datasets within data lakes is a critical challenge, particularly for semantically related tables that lack the explicit attributes needed to be joined. We study Discovery-Driven Integration, where the relevant sources and their missing relational structure must be discovered before integration. In this setting, unstructured text provides the evidence that connects otherwise disjoint tables. The fundamental challenge is to discover the relationships at a fine-grained level that connect individual rows from different tables through specific sentences. We formalize this task as Text-Mediated Join Path Discovery and propose a horizontal bidirectional cross-attention architecture called LOKI Latent-space Optimization for Knowledge Integration) that learns contextualized representations of table rows and sentences. Through a global table-text contrastive objective, fine-grained row-sentence associations emerge without explicit local supervision. Existing multi-modal discovery methods largely retrieve coarse-grained column-text associations, whereas integration systems assume supplied row-text links, schemas, or queries. LOKI instead transforms these implicit associations into explicit, interpretable join paths, organizes them into relation-consistent groups, and materializes them as typed integrated tables with sentence-level provenance. Comprehensive evaluations on real-world benchmarks demonstrate that LOKI consistently outperforms state-of-the-art multi-modal data discovery approaches, and materializes typed integrated tables with 0.982 macro typed-pair precision while being up to 40 times cheaper in LLM API cost than direct prompting.