REaR: Retrieve, Expand and Refine for Effective Multitable Retrieval

TL;DR

REAR is a three-stage, LLM-free framework that improves multi-table retrieval for complex question answering by separating relevance from structural compatibility, leading to better accuracy and efficiency.

Contribution

It introduces a novel three-stage retrieval process that enhances multi-table retrieval without relying on large language models, improving performance and scalability.

Findings

REAR improves retrieval quality on multiple datasets.

REAR achieves competitive performance with LLM-based systems.

The framework reduces latency and computational cost.

Abstract

Answering natural language queries over relational data often requires retrieving and reasoning over multiple tables, yet most retrievers optimize only for query-table relevance and ignore table table compatibility. We introduce REAR (Retrieve, Expand and Refine), a three-stage, LLM-free framework that separates semantic relevance from structural joinability for efficient, high-fidelity multi-table retrieval. REAR (i) retrieves query-aligned tables, (ii) expands these with structurally joinable tables via fast, precomputed column-embedding comparisons, and (iii) refines them by pruning noisy or weakly related candidates. Empirically, REAR is retriever-agnostic and consistently improves dense/sparse retrievers on complex table QA datasets (BIRD, MMQA, and Spider) by improving both multi-table retrieval quality and downstream SQL execution. Despite being LLM-free, it delivers performance competitive with state-of-the-art LLM-augmented retrieval systems (e.g.,ARM) while achieving much lower latency and cost. Ablations confirm complementary gains from expansion and refinement, underscoring REAR as a practical, scalable building block for table-based downstream tasks (e.g., Text-to-SQL).