Sentence-to-Label Generation Framework for Multi-task Learning of Japanese Sentence Classification and Named Entity Recognition

TL;DR

This paper presents a novel multi-task framework combining sentence classification and named entity recognition for Japanese, utilizing a sentence-to-label generation approach with a constraint mechanism to improve accuracy.

Contribution

The study introduces a unified generation-based framework for joint sentence classification and NER, demonstrating improved performance and format accuracy through a constraint mechanism.

Findings

SC accuracy increased by 1.13 points

NER accuracy increased by 1.06 points

Format accuracy improved from 63.61 to 100 with CM

Abstract

Information extraction(IE) is a crucial subfield within natural language processing. In this study, we introduce a Sentence Classification and Named Entity Recognition Multi-task (SCNM) approach that combines Sentence Classification (SC) and Named Entity Recognition (NER). We develop a Sentence-to-Label Generation (SLG) framework for SCNM and construct a Wikipedia dataset containing both SC and NER. Using a format converter, we unify input formats and employ a generative model to generate SC-labels, NER-labels, and associated text segments. We propose a Constraint Mechanism (CM) to improve generated format accuracy. Our results show SC accuracy increased by 1.13 points and NER by 1.06 points in SCNM compared to standalone tasks, with CM raising format accuracy from 63.61 to 100. The findings indicate mutual reinforcement effects between SC and NER, and integration enhances both tasks' performance.