GreekT5: A Series of Greek Sequence-to-Sequence Models for News Summarization

TL;DR

This paper introduces GreekT5, a series of transformer-based models tailored for Greek news summarization, demonstrating significant improvements over the existing GreekBART model in various evaluation metrics.

Contribution

The paper presents novel Greek-specific sequence-to-sequence models for news summarization, advancing NLP tools for low-resource Greek language processing.

Findings

Most models outperform GreekBART on evaluation metrics

Models achieve better summaries for Greek news articles

Evaluation code is publicly available for reproducibility

Abstract

Text summarization (TS) is a natural language processing (NLP) subtask pertaining to the automatic formulation of a concise and coherent summary that covers the major concepts and topics from one or multiple documents. Recent advancements in deep learning have led to the development of abstractive summarization transformer-based models, which outperform classical approaches. In any case, research in this field focuses on high resource languages such as English, while the corresponding work for low resource languages is still underdeveloped. Taking the above into account, this paper proposes a series of novel TS models for Greek news articles. The proposed models were thoroughly evaluated on the same dataset against GreekBART, which is the state-of-the-art model in Greek abstractive news summarization. Our evaluation results reveal that most of the proposed models significantly outperform GreekBART on various evaluation metrics. We make our evaluation code public, aiming to increase the reproducibility of this work and facilitate future research in the field.