Enhancing Temporal Understanding in LLMs for Semi-structured Tables

TL;DR

This paper analyzes the limitations of large language models in temporal reasoning over tabular data, introduces a new dataset and a novel method to improve their performance, and demonstrates significant enhancements in evidence-based reasoning.

Contribution

It presents TempTabQA dataset, C.L.E.A.R method, and insights for improving LLM temporal reasoning over semi-structured tables, advancing the state of the art.

Findings

C.L.E.A.R significantly improves reasoning accuracy.

Auxiliary data boosts model performance.

Enhanced models outperform baselines in temporal tasks.

Abstract

Temporal reasoning over tabular data presents substantial challenges for large language models (LLMs), as evidenced by recent research. In this study, we conduct a comprehensive analysis of temporal datasets to pinpoint the specific limitations of LLMs. Our investigation leads to enhancements in TempTabQA, a dataset specifically designed for tabular temporal question answering. We provide critical insights for improving LLM performance in temporal reasoning tasks with tabular data. Furthermore, we introduce a novel approach, C.L.E.A.R to strengthen LLM capabilities in this domain. Our findings demonstrate that our method significantly improves evidence-based reasoning across various models. Additionally, our experimental results reveal that indirect supervision with auxiliary data substantially boosts model performance in these tasks. This work contributes to a deeper understanding of LLMs' temporal reasoning abilities over tabular data and promotes advancements in their application across diverse fields.