Research map / Label-efficient and robust learning / Recognition and long-tail
Learning to Extract Coherent Summary via Deep Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the coherence reward to RNES (ablation) | RNES · CNN/Daily Mail test | ROUGE-1 F1 (%): 41.25 -> 40.95 | No clear effect |
| Add the coherence reward to RNES (ablation) | RNES · CNN/Daily Mail test | ROUGE-2 F1 (%): 18.87 -> 18.63 | No clear effect |
| Add the coherence reward to RNES (ablation) | RNES · CNN/Daily Mail test | ROUGE-L F1 (%): 37.75 -> 37.41 | No clear effect |
| Add the coherence reward to RNES (ablation) | RNES · Human evaluation, 50 test documents | Coherence rank: 1.325 -> 1.092 | Helped |
| Add the coherence reward to RNES (ablation) | RNES · Human evaluation, 50 test documents | Informativeness rank: 1.183 -> 1.125 | Helped |
| Add the coherence reward to RNES (ablation) | RNES · Human evaluation, 50 test documents | Overall rank: 1.492 -> 1.209 | Helped |
| Set the coherence weight to 1.0 or 0.1 (ablation) | RNES · CNN/Daily Mail | ROUGE | Hurt |
| Set the coherence weight to 0.005 (ablation) | RNES · CNN/Daily Mail | Coherence reward: 0.005 | Hurt |
| Set the coherence weight to 0.01 (ablation) | RNES · CNN/Daily Mail | Coherence and ROUGE rewards: 0.01 | Helped |
Closest papers by meaning
- A Deep Reinforced Model for Abstractive Summarization (ICLR 2018 · similarity 0.87)
- Controlling the Amount of Verbatim Copying in Abstractive Summarization (AAAI 2020 · similarity 0.87)
- SummaRuNNer: A Recurrent Neural Network Based Sequence Model for Extractive Summarization of Documents (AAAI 2017 · similarity 0.87)
- Faithful to the Original: Fact Aware Neural Abstractive Summarization (AAAI 2018 · similarity 0.86)
- BBScore: A Brownian Bridge Based Metric for Assessing Text Coherence (AAAI 2024 · similarity 0.85)
- Learning to Summarize by Learning to Quiz: Adversarial Agentic Collaboration for Long Document Summarization (ICLR 2026 · similarity 0.84)
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.