用分解-清洗-推理三步法提升大模型处理表格数值推理能力
TabDSR: Decompose, Sanitize, and Reason for Complex Numerical Reasoning in Tabular Data
- 将复杂问题拆解，清洗噪声表格，生成可执行代码推理
- 在TAT-QA等数据集上准确率提升最高达19.87%
- 适合需要精准表格分析的金融、科研场景使用
表格数据中的复杂推理在现实数据分析中至关重要，但大语言模型常因查询复杂、数据噪声和数值能力有限而表现不佳。为此，我们提出TabDSR框架，包含：(1) 查询分解器，将复杂问题拆解；(2) 表格清洗器，清理并过滤噪声表格；(3) 基于程序思维（PoT）的推理器，生成可执行代码从清洗后的表格中推导答案。为确保评估无偏且防止数据泄露，我们构建了新数据集CalTab151，专用于表格数值推理。实验表明，TabDSR持续优于现有方法，在TAT-QA、TableBench和TabDSR上分别实现8.79%、6.08%和19.87%的准确率提升。该框架可无缝集成主流LLM，为复杂表格数值推理提供稳健解决方案。
原文摘要 · Abstract (English)
Complex reasoning over tabular data is crucial in real-world data analysis, yet large language models (LLMs) often underperform due to complex queries, noisy data, and limited numerical capabilities. To address these issues, we propose TabDSR, a framework consisting of: (1) a query decomposer that breaks down complex questions, (2) a table sanitizer that cleans and filters noisy tables, and (3) a program-of-thoughts (PoT)-based reasoner that generates executable code to derive the final answer from the sanitized table. To ensure unbiased evaluation and mitigate data leakage, we introduce a new dataset, CalTab151, specifically designed for complex numerical reasoning over tables. Experimental results demonstrate that TabDSR consistently outperforms existing methods, achieving state-of-the-art (SOTA) performance with 8.79%, 6.08%, and 19.87% accuracy improvement on TAT-QA, TableBench, and TabDSR, respectively. Moreover, our framework integrates seamlessly with mainstream LLMs, providing a robust solution for complex tabular numerical reasoning. These findings highlight the effectiveness of our framework in enhancing LLM performance for complex tabular numerical reasoning. Data and code are available upon request.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。