Is Table Retrieval a Solved Problem? Exploring Join-Aware Multi-Table Retrieval

2024-04-15

Annual Meeting of the Association for Computational Linguistics

AI Summary

To address the challenges of multi-table retrieval and implicit cross-table semantic associations in open-domain table-based question answering, this paper proposes the first join-aware table retrieval framework, which explicitly models inter-table join relationships during retrieval. Unlike single-table matching or query decomposition paradigms, our approach integrates join relation inference into the ranking process via mixed-integer programming (MIP), jointly optimizing table–query relevance and table–table join strength. This enables synergistic modeling of both semantic and structural relevance in multi-table re-ranking. Experiments demonstrate that our method achieves a 9.3% improvement in table retrieval F1 score and a 5.4% gain in end-to-end question answering accuracy, significantly outperforming existing state-of-the-art methods.

Technology Category

• Data Mining & Knowledge Management: Intelligent Query Processing
• Natural Language Processing: Question Answering
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

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

Abstract

Retrieving relevant tables containing the necessary information to accurately answer a given question over tables is critical to open-domain question-answering (QA) systems. Previous methods assume the answer to such a question can be found either in a single table or multiple tables identified through question decomposition or rewriting. However, neither of these approaches is sufficient, as many questions require retrieving multiple tables and joining them through a join plan that cannot be discerned from the user query itself. If the join plan is not considered in the retrieval stage, the subsequent steps of reasoning and answering based on those retrieved tables are likely to be incorrect. To address this problem, we introduce a method that uncovers useful join relations for any query and database during table retrieval. We use a novel re-ranking method formulated as a mixed-integer program that considers not only table-query relevance but also table-table relevance that requires inferring join relationships. Our method outperforms the state-of-the-art approaches for table retrieval by up to 9.3% in F1 score and for end-to-end QA by up to 5.4% in accuracy.

Problem

• Multi-table Querying
• Tabular Data Understanding
• Cross-table Relationship Analysis

Innovation

• Multi-table Information Retrieval
• Implicit Relationship Identification
• Accuracy Improvement