Transformer-Based Models for Automatic Identification of Argument Relations: A Cross-Domain Evaluation

TL;DR

This paper evaluates transformer-based models for argument relation identification across multiple domains, demonstrating their effectiveness and domain dependence using the large US2016 debate corpus.

Contribution

It provides a comprehensive analysis of transformer models in argument mining and assesses their cross-domain generalization capabilities.

Findings

Transformer models achieve up to 0.70 macro F1-score on US2016 corpus.

Models show reduced performance (0.61 macro F1-score) in cross-domain evaluation.

The study highlights domain dependence in argument relation prediction.

Abstract

Argument Mining is defined as the task of automatically identifying and extracting argumentative components (e.g., premises, claims, etc.) and detecting the existing relations among them (i.e., support, attack, rephrase, no relation). One of the main issues when approaching this problem is the lack of data, and the size of the publicly available corpora. In this work, we use the recently annotated US2016 debate corpus. US2016 is the largest existing argument annotated corpus, which allows exploring the benefits of the most recent advances in Natural Language Processing in a complex domain like Argument (relation) Mining. We present an exhaustive analysis of the behavior of transformer-based models (i.e., BERT, XLNET, RoBERTa, DistilBERT and ALBERT) when predicting argument relations. Finally, we evaluate the models in five different domains, with the objective of finding the less domain dependent model. We obtain a macro F1-score of 0.70 with the US2016 evaluation corpus, and a macro F1-score of 0.61 with the Moral Maze cross-domain corpus.