Research map / Large language models / Language, safety and interpretability
Contextualized Rewriting for Text Summarization
Guangsheng Bao, Yue Zhang
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add contextualized rewriting to BERT-Ext | BERT-Ext · CNN/Daily Mail | ROUGE-1: 41.04 -> 43.52 (+2.48) | Improved |
| Add contextualized rewriting to BERT-Ext | BERT-Ext · CNN/Daily Mail | ROUGE-2: 19.56 -> 20.57 (+1.01) | Improved |
| Add contextualized rewriting to BERT-Ext | BERT-Ext · CNN/Daily Mail | ROUGE-L: 37.66 -> 40.56 (+2.90) | Improved |
| Add group-tag-guided contextual document input instead of plain abstractive generation (ablation) | BERT-Abs · CNN/Daily Mail | ROUGE-1: 41.70 -> 43.52 | Improved |
| Add group-tag-guided contextual document input instead of plain abstractive generation (ablation) | BERT-Abs · CNN/Daily Mail | ROUGE-2: 19.06 -> 20.57 | Improved |
| Add group-tag-guided contextual document input instead of plain abstractive generation (ablation) | BERT-Abs · CNN/Daily Mail | ROUGE-L: 38.71 -> 40.56 | Improved |
| Add contextualized rewriting to BERT-Hybrid | BERT-Hybrid · CNN/Daily Mail | ROUGE-1: +1.76 | Improved |
| Add contextualized rewriting to BERT-Hybrid | BERT-Hybrid · CNN/Daily Mail | ROUGE-2: +1.26 | Improved |
| Add contextualized rewriting to BERT-Hybrid | BERT-Hybrid · CNN/Daily Mail | ROUGE-L: +1.7 | Improved |
| Add contextualized rewriting to LEAD-3 | LEAD-3 · CNN/Daily Mail | ROUGE-L: +1.49 | Improved |
| Add contextualized rewriting to BERTSUMEXT without trigram blocking | BERTSUMEXT w/o Tri-Bloc · CNN/Daily Mail | ROUGE-L: +1.42 | Improved |
| Remove trigram blocking from BERTSUMEXT (ablation) | BERTSUMEXT · CNN/Daily Mail | ROUGE | Hurt |
Closest papers by meaning
- BooookScore: A systematic exploration of book-length summarization in the era of LLMs (ICLR 2024 · similarity 0.83)
- Document Summarization with Conformal Importance Guarantees (NeurIPS 2025 · similarity 0.83)
- RewriteLM: An Instruction-Tuned Large Language Model for Text Rewriting (AAAI 2024 · similarity 0.82)
- Language Models of Code Are Few-Shot Planners and Reasoners for Multi-Document Summarization with Attribution (AAAI 2025 · similarity 0.82)
- MaFeRw: Query Rewriting with Multi-Aspect Feedbacks for Retrieval-Augmented Large Language Models (AAAI 2025 · similarity 0.82)
- Unsupervised Extractive Summarization with Learnable Length Control Strategies (AAAI 2024 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
