Function-words Enhanced Attention Networks for Few-Shot Inverse Relation Classification

TL;DR

This paper introduces FAEA, a novel attention framework leveraging function words and meta-learning to improve few-shot inverse relation classification, significantly enhancing accuracy in low-data scenarios.

Contribution

The paper proposes a function words enhanced attention model with adaptive message passing for few-shot inverse relation classification, addressing intra-class redundancy and inter-class differences.

Findings

FAEA outperforms baselines in inverse relation accuracy

Achieves 14.33% improvement in 1-shot setting on FewRel1.0

Effective reduction of intra-class redundancy through message passing

Abstract

The relation classification is to identify semantic relations between two entities in a given text. While existing models perform well for classifying inverse relations with large datasets, their performance is significantly reduced for few-shot learning. In this paper, we propose a function words adaptively enhanced attention framework (FAEA) for few-shot inverse relation classification, in which a hybrid attention model is designed to attend class-related function words based on meta-learning. As the involvement of function words brings in significant intra-class redundancy, an adaptive message passing mechanism is introduced to capture and transfer inter-class differences.We mathematically analyze the negative impact of function words from dot-product measurement, which explains why message passing mechanism effectively reduces the impact. Our experimental results show that FAEA outperforms strong baselines, especially the inverse relation accuracy is improved by 14.33% under 1-shot setting in FewRel1.0.