Character-Level Models versus Morphology in Semantic Role Labeling

TL;DR

This paper compares character-level, word, and morphological models for semantic role labeling across multiple languages, revealing strengths and limitations of character models in capturing morphological structure and handling various linguistic challenges.

Contribution

It provides a comprehensive analysis of character-level models versus morphology-based models in SRL, highlighting their respective advantages and limitations across languages.

Findings

Character models perform well but struggle with morphological complexity.

Morphology-aware models improve accuracy in morphologically rich languages.

Character models are sensitive to out-of-domain data and training data size.

Abstract

Character-level models have become a popular approach specially for their accessibility and ability to handle unseen data. However, little is known on their ability to reveal the underlying morphological structure of a word, which is a crucial skill for high-level semantic analysis tasks, such as semantic role labeling (SRL). In this work, we train various types of SRL models that use word, character and morphology level information and analyze how performance of characters compare to words and morphology for several languages. We conduct an in-depth error analysis for each morphological typology and analyze the strengths and limitations of character-level models that relate to out-of-domain data, training data size, long range dependencies and model complexity. Our exhaustive analyses shed light on important characteristics of character-level models and their semantic capability.