Interpretable LLM-based Table Question Answering

2024-12-16

arXiv.org

AI Summary

In high-stakes domains such as finance and healthcare, large language models (LLMs) exhibit ambiguous explanations and untrustworthy decision-making when performing table-based question answering (Table QA). Method: This paper proposes Plan-of-SQLs (POS), an interpretable framework that explicitly generates multi-step SQL reasoning plans to construct structured, verifiable reasoning chains. Contribution/Results: POS introduces the first principled explanation paradigm for Table QA, demonstrating 90% agreement between LLMs and human annotators in explanation quality evaluation—enabling LLMs to replace manual assessment efficiently. It achieves state-of-the-art or leading accuracy on TabFact, WikiTQ, and FetaQA, reduces LLM invocation costs by over 40%, exhibits strong robustness on large tables, and delivers optimal explanation quality validated through both human and LLM evaluation pathways.

Technology Category

• Natural Language Processing: Question Answering
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Interpretability for Table Question Answering (Table QA) is critical, particularly in high-stakes industries like finance or healthcare. Although recent approaches using Large Language Models (LLMs) have significantly improved Table QA performance, their explanations for how the answers are generated are ambiguous. To fill this gap, we introduce Plan-of-SQLs (POS), an interpretable Table QA approach designed to improve users' understanding of model decision-making. Through qualitative and quantitative evaluations with human and LLM judges, we show that: First, POS is the highest-quality explanation method, helps human users understand model behaviors, and facilitates model prediction verification. Second, when evaluated on popular and standard Table QA datasets (TabFact, WikiTQ, and FetaQA), POS achieves QA accuracy that is competitive with or superior to existing methods, while also offering greater efficiency-requiring significantly fewer LLM calls and table database queries-and robust performance on large-sized tables. Finally, we observe high agreement (up to 90%) between LLMs and human users when making decisions based on the same explanations, suggesting that LLMs could serve as an effective proxy for humans in evaluating explanations. This finding enables faster, more affordable evaluation of AI explanations-possibly accelerating trustworthy AI research while maintaining reliable judgments on interpretability.

Problem

• Enhancing interpretability in LLM-based Table QA systems
• Improving explanation quality for model decision-making
• Achieving competitive QA accuracy with efficient resource usage

Innovation

• Plan-of-SQLs enhances interpretability in Table QA
• POS reduces LLM calls and database queries efficiently
• LLMs and humans show high agreement on explanations