Hybrid text summarization: Integrating extractive and abstractive models for enhanced cross-domain summarization

Abstract

This paper investigates the effectiveness of various text summarization models—extractive, abstractive, generative, and a proposed hybrid model—across three distinct content domains: news articles, academic papers, and social media. The study evaluates the performance of these models using standard metrics, including ROUGE, BLEU, and METEOR, providing a comprehensive assessment of each model’s strengths and limitations.

Results indicate that the extractive model excels in factual retention for news summaries, while the abstractive model demonstrates superior coherence and conciseness in academic papers. The generative model effectively captures contextual nuances in social media content, albeit with some loss in factual precision. Notably, the proposed hybrid model achieves the highest overall performance by integrating the strengths of extractive and abstractive techniques, enhancing adaptability across different domains.

These findings highlight the importance of selecting appropriate summarization strategies tailored to specific content types and demonstrate the potential of hybrid approaches in improving summary relevance and quality. Future research should explore further optimizations, including domain adaptation techniques and reinforcement learning strategies, to refine hybrid summarization methods.

Keywords

text summarization hybrid model ROUGE scores BLEU score METEOR score domain-Specific summarization

Get full access to this article

View all access options for this article.

References

Burchardt

Ochoa

, et al. Extractive and abstractive summarization in social media: a review of recent approaches. In: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics pp.550–561, 2020.

Shafiei Bajestani

, et al. Cross-domain summarization: A survey. arXiv preprint arXiv:2002.04604 2020.

Liu

Lapata

. Text summarization with pretrained encoders. arXiv preprint arXiv:1903.10318 2019.

Devlin

Chang

M-W

, et al. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018.

Vaswani

Shard

, et al. Attention is all you need. Adv Neural Inf Process Syst 2017; 30.

Khandelwal

, et al. Generalization through memorization: Nearest neighbor language models. In: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics pp.2755–2764, 2020.

Zhang

, et al. Pre-training transformers as energy-based cloze models. arXiv preprint arXiv:1901.09264 2019.

Ganin

Lempitsky

. Domain-adversarial training of neural networks. In: Proceedings of the 33rd International Conference on Machine Learning pp.1180–1189, 2016.

Kurland

Schutze

. A kernel-based approach to text summarization. In: Proceedings of the 28th annual international ACM SIGIR conference on research and development in information retrieval, pp.247–254, 2005.

10.

Pan

Yang

. A survey on transfer learning. IEEE Trans Knowl Data Eng 2009; 22: 1345–1359.

11.

Sun

et al. Deep transfer learning for named entity recognition. ACM Trans Intell Syst Technol 2021; 12.

12.

Howard

Ruder

. Universal language model fine-tuning for text classification. ACL 2018 2018.

13.

Hsu

C-H

et al. Language models are few-shot learners. arXiv preprint arXiv:2006.05529 2020.

14.

Nallapati

, et al. Abstractive text summarization using sequence-to-sequence rnns and beyond. In: Proceedings of the 20th sign language processing conference pp.280–290, 2016.

15.

See

Liu

Manning

. Get to the point: Summarization with pointer-generator networks. In: Proceedings of the 55th annual meeting of the association for computational linguistics pp.421–430, 2017.

16.

et al. Reinforcement learning for abstractive text summarization. In: Proceedings of the 55th annual meeting of the association for computational linguistics pp.1768–1778, 2017.

17.

Zhou

, et al. Hierarchical attention mechanism for text summarization. In: Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies pp.1641–1650, 2019.

18.

Hermann

Kocisky

, et al. Teaching machines to read and comprehend. In: Advances in neural information processing systems pp.1684–1692, 2015.

19.

Cohan

Goharian

, et al. A discourse-aware attention model for abstractive summarization of academic papers. In: Proceedings of the 27th international conference on computational linguistics pp.1667–1678, 2018.

20.

Paszke

Gross

Wang

, et al. Pytorch: An imperative style, high-performance deep learning library. In: Advances in neural information processing systems pp.8024–8035, 2019.

21.

Wolf Chaumond

Debut

Sanh

, et al. Transformers: State-of-the-art natural language processing. In: Proceedings of the 2020 conference on empirical methods in natural language processing: system demonstrations pp.38–45, 2020.

22.

Kingma

. Adam: a method for stochastic optimization. In: Proceedings of the 3rd international conference for learning representations (ICLR). https://arxiv.org/abs/1412.6980, 2014.

23.

Srivastava

Hinton

Krizhevsky

, et al. Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 2014; 15: 1929–1958.