Getting the Most out of Simile Recognition

TL;DR

This paper introduces HGSR, a novel model for simile recognition that uses expressive features like POS tags, dependency trees, and decoding interdependence, significantly improving accuracy over existing methods.

Contribution

The paper proposes a new model, HGSR, which integrates heterogeneous graph features and decoding interdependence for more effective simile recognition.

Findings

HGSR outperforms state-of-the-art systems

Expressive features improve recognition accuracy

Decoding interdependence enhances model performance

Abstract

Simile recognition involves two subtasks: simile sentence classification that discriminates whether a sentence contains simile, and simile component extraction that locates the corresponding objects (i.e., tenors and vehicles). Recent work ignores features other than surface strings. In this paper, we explore expressive features for this task to achieve more effective data utilization. Particularly, we study two types of features: 1) input-side features that include POS tags, dependency trees and word definitions, and 2) decoding features that capture the interdependence among various decoding decisions. We further construct a model named HGSR, which merges the input-side features as a heterogeneous graph and leverages decoding features via distillation. Experiments show that HGSR significantly outperforms the current state-of-the-art systems and carefully designed baselines, verifying the effectiveness of introduced features. Our code is available at https://github.com/DeepLearnXMU/HGSR.