Long-Document Cross-Lingual Summarization

TL;DR

This paper introduces Perseus, a new dataset for long-document cross-lingual summarization, and evaluates baseline models, highlighting the advantages of end-to-end methods over pipeline approaches.

Contribution

It provides the first long-document CLS dataset and benchmarks, facilitating future research on complex, lengthy texts across languages.

Findings

End-to-end models outperform pipeline models on Perseus.

Long documents pose unique challenges for CLS.

Perseus dataset contains 94K Chinese documents with English summaries.

Abstract

Cross-Lingual Summarization (CLS) aims at generating summaries in one language for the given documents in another language. CLS has attracted wide research attention due to its practical significance in the multi-lingual world. Though great contributions have been made, existing CLS works typically focus on short documents, such as news articles, short dialogues and guides. Different from these short texts, long documents such as academic articles and business reports usually discuss complicated subjects and consist of thousands of words, making them non-trivial to process and summarize. To promote CLS research on long documents, we construct Perseus, the first long-document CLS dataset which collects about 94K Chinese scientific documents paired with English summaries. The average length of documents in Perseus is more than two thousand tokens. As a preliminary study on long-document CLS, we build and evaluate various CLS baselines, including pipeline and end-to-end methods. Experimental results on Perseus show the superiority of the end-to-end baseline, outperforming the strong pipeline models equipped with sophisticated machine translation systems. Furthermore, to provide a deeper understanding, we manually analyze the model outputs and discuss specific challenges faced by current approaches. We hope that our work could benchmark long-document CLS and benefit future studies.