A BART-based approach with hierarchical strategy for Vietnamese abstractive multi-document summarization

2026-06-17 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Advanced, quick

Summary

A technical report introduces a BART-based hierarchical strategy for Vietnamese abstractive multi-document summarization, addressing a challenge from the International Workshop on Vietnamese Language and Speech Processing (VLSP) 2022. The approach condenses individual documents before aggregation and final summarization. A novel document shortening strategy, guided by the "golden summary," ensures strong correlation across the hierarchical stages. This method achieved a ROUGE2-F1 score of 0.2468 on VLSP's public test set, producing fluent and concise summaries. Additionally, the authors utilized external sources to augment data quantity for Vietnamese multi-document summarization, making this supplementary data available to the community.

Key takeaway

For NLP Engineers developing multi-document summarization systems for low-resource languages like Vietnamese, this research offers a validated hierarchical BART-based approach. You should consider implementing a golden summary-driven document shortening strategy to improve inter-stage correlation and utilize external data sources to enhance model performance. This method achieved a ROUGE2-F1 of 0.2468, suggesting a robust framework for producing fluent and concise summaries.

Key insights

A BART-based hierarchical strategy with golden summary-driven shortening improves Vietnamese multi-document abstractive summarization, achieving 0.2468 ROUGE2-F1.

Principles

• Hierarchical summarization condenses documents then aggregates.
• Golden summary-driven shortening enhances stage correlation.
• External data augmentation improves summarization performance.

Method

The method condenses each document using a novel golden summary-driven shortening strategy, followed by aggregation and abstractive summarization with a BART-based model.

In practice

• Utilize BART for abstractive summarization tasks.
• Implement hierarchical strategies for multi-document inputs.
• Incorporate external data to boost training sets.

Topics

• Multi-document Summarization
• Abstractive Summarization
• Vietnamese NLP
• BART Model
• Hierarchical Summarization
• VLSP 2022

Best for: Research Scientist, AI Scientist, NLP Engineer

Related on AIssential

• MASF: A Multi-Model Adaptive Selection Framework for Abstractive Text summarization — Combining multiple fine-tuned transformer models with an adaptive selection mechanism significantly improves abstractive summarization quality.
• Evaluation of Chunking Strategies for Effective Text Embedding in Low-Resource Language on Agricultural Documents — Effective text chunking significantly improves RAG performance in low-resource, morphologically complex languages.
• A Training-Free Mixture-of-Agents Framework for Multi-Document Summarization using LLMs and Knowledge Graphs — A training-free agent framework combines LLMs and knowledge graphs for effective multi-document summarization.
• A Comparative Study of Multi-document Summarization Techniques — The paper compares multi-document summarization techniques, contributing to NLP research.
• MiqraBERT: Regression-Based Sentence-BERT Finetuning for Biblical Hebrew Parallel Detection — Finetuning a BERT model with regression for semantic similarity effectively detects textual reuse in ancient languages beyond lexical overlap.
• The Essential Guide to Effectively Summarizing Massive Documents, Part 2 — Clustering and representative chunk selection enable scalable, structured summarization of large documents, significantly reducing token load.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.