CGPT: Cluster-Guided Partial Tables with LLM-Generated Supervision for Table Retrieval

2026-01-22

AI Summary

This work addresses the limitations of general-purpose embedding models in table retrieval, which stem from structural semantic compression and mismatches between queries and tabular content. To overcome these challenges, the authors propose a clustering-guided partial table construction approach that leverages K-means clustering and large language models (LLMs) to generate semantically diverse synthetic queries, thereby creating high-quality supervision signals. The embedding model is then fine-tuned using contrastive learning with hard negative examples. This strategy substantially enhances semantic coverage and cross-domain generalization, achieving an average 16.54% improvement in Recall@1 across four public benchmarks—outperforming existing methods. Notably, the approach remains highly effective even when using smaller-scale LLMs and a unified multi-domain corpus.

Technology Category

• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

General-purpose embedding models have demonstrated strong performance in text retrieval but remain suboptimal for table retrieval, where highly structured content leads to semantic compression and query-table mismatch. Recent LLM-based retrieval augmentation methods mitigate this issue by generating synthetic queries, yet they often rely on heuristic partial-table selection and seldom leverage these synthetic queries as supervision to improve the embedding model. We introduce CGPT, a training framework that enhances table retrieval through LLM-generated supervision. CGPT constructs semantically diverse partial tables by clustering table instances using K-means and sampling across clusters to broaden semantic coverage. An LLM then generates synthetic queries for these partial tables, which are used in hard-negative contrastive fine-tuning to refine the embedding model. Experiments across four public benchmarks (MimoTable, OTTQA, FetaQA, and E2E-WTQ) show that CGPT consistently outperforms retrieval baselines, including QGpT, with an average R@1 improvement of 16.54 percent. In a unified multi-domain corpus setting, CGPT further demonstrates strong cross-domain generalization and remains effective even when using smaller LLMs for synthetic query generation. These results indicate that semantically guided partial-table construction, combined with contrastive training from LLM-generated supervision, provides an effective and scalable paradigm for large-scale table retrieval. Our code is available at https://github.com/yumeow0122/CGPT.

Problem

• table retrieval
• semantic compression
• query-table mismatch
• structured data
• embedding models

Innovation

• table retrieval
• LLM-generated supervision
• partial table construction
• contrastive fine-tuning
• semantic clustering