TempTabQA: Temporal Question Answering for Semi-Structured Tables

TL;DR

TempTabQA introduces a new dataset and task for evaluating how well NLP models can perform temporal reasoning on semi-structured tables, highlighting current limitations and guiding future improvements.

Contribution

The paper presents TempTabQA, a large dataset for temporal question answering on semi-structured tables, and evaluates state-of-the-art models, revealing significant gaps in temporal reasoning abilities.

Findings

Top models lag human performance by 13.5 F1 points.

Existing models struggle with temporal reasoning in semi-structured data.

TempTabQA serves as a challenging benchmark for future research.

Abstract

Semi-structured data, such as Infobox tables, often include temporal information about entities, either implicitly or explicitly. Can current NLP systems reason about such information in semi-structured tables? To tackle this question, we introduce the task of temporal question answering on semi-structured tables. We present a dataset, TempTabQA, which comprises 11,454 question-answer pairs extracted from 1,208 Wikipedia Infobox tables spanning more than 90 distinct domains. Using this dataset, we evaluate several state-of-the-art models for temporal reasoning. We observe that even the top-performing LLMs lag behind human performance by more than 13.5 F1 points. Given these results, our dataset has the potential to serve as a challenging benchmark to improve the temporal reasoning capabilities of NLP models.