Part & Whole Extraction: Towards A Deep Understanding of Quantitative Facts for Percentages in Text

TL;DR

This paper presents a novel sequence tagging approach with a skip mechanism to extract quantitative part-whole facts related to percentages in text, enhancing understanding for applications like infographic generation.

Contribution

It introduces a skip-based sequence tagging model specifically designed for extracting part and whole entities associated with percentages, improving accuracy over existing methods.

Findings

Effective in extracting part-whole relationships for percentages

Improved performance on sequence tagging and NER tasks

Skip mechanism enhances extraction accuracy

Abstract

We study the problem of quantitative facts extraction for text with percentages. For example, given the sentence "30 percent of Americans like watching football, while 20% prefer to watch NBA.", our goal is to obtain a deep understanding of the percentage numbers ("30 percent" and "20%") by extracting their quantitative facts: part ("like watching football" and "prefer to watch NBA") and whole ("Americans). These quantitative facts can empower new applications like automated infographic generation. We formulate part and whole extraction as a sequence tagging problem. Due to the large gap between part/whole and its corresponding percentage, we introduce skip mechanism in sequence modeling, and achieved improved performance on both our task and the CoNLL-2003 named entity recognition task. Experimental results demonstrate that learning to skip in sequence tagging is promising.