Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper addresses the challenge of Vietnamese multi-document abstractive summarization by employing a hierarchical strategy that condenses individual documents before aggregation. The authors introduce a novel document shortening method guided by a golden summary, ensuring high relevance across the summarization stages. Their approach achieves a ROUGE2-F1 score of 0.2468 on the VLSP public test set, demonstrating the effectiveness of their method in producing fluent and concise summaries while also enhancing the dataset available for the community.
A novel document shortening strategy driven by golden summaries significantly boosts the quality of Vietnamese multi-document summarization.
In this technical report, we focus on solving the challenge of Vietnamese multi-document abstractive summarization, introduced in the International Workshop on Vietnamese Language and Speech Processing (VLSP) 2022. We choose to follow the popular hierarchical approach, i.e. condensing each document followed by aggregation and summarization. We propose a novel yet simple strategy to shorten documents that is driven by the golden summary, thus ensuring high correlation between stages of the hierarchical approach. Our method achieves a ROUGE2-F1 score of 0.2468 on the VLSP's public test set, and can produce fluent and concise summaries. Additionally, we utilize external sources for extra data, which greatly enhances the quantity of data for Vietnamese multi-document summarization. The additional data is made available for the community.