Comparison between parameter-efficient techniques and full fine-tuning: A case study on multilingual news article classification

TL;DR

This study compares parameter-efficient fine-tuning methods like Adapters and LoRA with full fine-tuning for multilingual news article classification, analyzing their performance and computational costs across various languages and data scenarios.

Contribution

It provides a comprehensive analysis of how parameter-efficient techniques perform on complex multilingual classification tasks compared to full fine-tuning.

Findings

Parameter-efficient methods can match or outperform full fine-tuning in certain multilingual tasks.

Efficiency gains are notable in low-resource and multilabel classification scenarios.

Performance varies across languages and input lengths, highlighting context-dependent effectiveness.

Abstract

Adapters and Low-Rank Adaptation (LoRA) are parameter-efficient fine-tuning techniques designed to make the training of language models more efficient. Previous results demonstrated that these methods can even improve performance on some classification tasks. This paper complements the existing research by investigating how these techniques influence the classification performance and computation costs compared to full fine-tuning when applied to multilingual text classification tasks (genre, framing, and persuasion techniques detection; with different input lengths, number of predicted classes and classification difficulty), some of which have limited training data. In addition, we conduct in-depth analyses of their efficacy across different training scenarios (training on the original multilingual data; on the translations into English; and on a subset of English-only data) and different languages. Our findings provide valuable insights into the applicability of the parameter-efficient fine-tuning techniques, particularly to complex multilingual and multilabel classification tasks.