arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2212.10097cs.CLcs.AIcs.DB
收藏

基于自训练框架的无监督复杂表格推理优化技术

Optimization Techniques for Unsupervised Complex Table Reasoning via Self-Training Framework

Tsinghua University(清华大学);University of Chinese Academy of Sciences(中国科学院大学)

Zhenyu Li, Xiuxing Li, Sunqi Fan, Jianyong Wang

AI总结:

针对复杂表格推理标注数据不足的问题,提出含程序管理、程序转换等模块的无监督自训练框架UCTR-ST,性能达监督模型90%以上,还可作数据增强提升低资源领域监督模型表现。

摘要:

结构化表格数据是众多领域中的基础数据类型,而表格推理能力对于回答问题和验证假设至关重要。然而,为复杂推理任务构建标注数据需要大量人力,且现有标注数据的数量仍不足以支撑现实应用的复杂需求。为解决标注不足的挑战,我们提出了一种用于无监督复杂表格推理的自训练框架UCTR-ST,该框架可生成具有复杂逻辑的多样化合成数据。具体而言,UCTR-ST包含若干关键技术:我们基于“程序管理”组件聚合多样化程序并在表格上执行这些程序,还通过功能强大的“程序转换”模块弥合程序与文本之间的差距,生成具有复杂逻辑的自然语言句子。此外,我们使用“表格-文本操纵器”优化流程,以处理表格-文本联合推理场景。整个框架利用自训练技术来利用未标注的训练数据,在现实数据上测试时实现了显著的性能提升。实验结果表明,UCTR-ST在不同任务和领域上达到了有监督模型性能的90%以上,降低了对人工标注的依赖。此外,我们的方法可作为一种数据增强技术,显著提升低资源领域中有监督模型的性能。

英文摘要:

Structured tabular data is a fundamental data type in numerous fields, and the capacity to reason over tables is crucial for answering questions and validating hypotheses. However, constructing labeled data for complex reasoning tasks is labor intensive, and the quantity of annotated data remains insufficient to support the intricate demands of real-world applications. To address the insufficient annotation challenge, we present a self-training framework for unsupervised complex tabular reasoning (UCTR-ST) by generating diverse synthetic data with complex logic. Specifically, UCTR-ST incorporates several essential techniques: we aggregate diverse programs and execute them on tables based on a "Program-Management" component, and we bridge the gap between programs and text with a powerful "Program-Transformation" module that generates natural language sentences with complex logic. Furthermore, we optimize the procedure using a "Table-Text Manipulator" to handle joint table-text reasoning scenarios. The entire framework utilizes self-training techniques to leverage the unlabeled training data, which results in significant performance improvements when tested on real-world data. Experimental results demonstrate that UCTRST achieves above 90% of the supervised model performance on different tasks and domains, reducing the dependence on manual annotation. Additionally, our approach can serve as a data augmentation technique, significantly boosting the performance of supervised models in low-resourced domains.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑