Research map / Image generation / Autoregressive and latent generation
Attractive or Faithful? Popularity-Reinforced Learning for Inspired Headline Generation
Yun-Zhu Song, Hong-Han Shuai, Sung-Lin Yeh, Yi-Lun Wu, Lun-Wei Ku, Wen-Chih Peng
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| use PORL-HG instead of Chen et al. | Headline generation · Headline test set | Attractiveness: 34.53 -> 44.06 | Improved |
| use PORL-HG instead of Chen et al. | Headline generation · Headline test set | ROUGE-L: 30.26 -> 29.48 | Hurt |
| add popularity and topic losses (ablation) | PORL-HG without popularity and topic losses · Headline test set | Attractiveness: 40.76 -> 44.06 | Improved |
| add popularity and topic losses (ablation) | PORL-HG without popularity and topic losses · Headline test set | Meteor: 17.25 -> 18.07 | Improved |
| add popularity loss (ablation) | PORL-HG without popularity loss · Headline test set | Attractiveness: 40.96 -> 44.06 | Improved |
| add popularity loss (ablation) | PORL-HG without popularity loss · Headline test set | Meteor: 17.53 -> 18.07 | Improved |
| add PTA (ablation) | PORL-HG without PTA · Headline test set | Attractiveness: 40.76 -> 44.06 | Improved |
| use PORL-HG instead of Chen et al. | Headline generation, human evaluation · User study | Attractiveness preference: 36.90% -> 63.10% | Improved |
| use PORL-HG instead of Chen et al. | Headline generation, human evaluation · User study | Relevance selections: 34.26% -> 35.65% | No clear effect |
Closest papers by meaning
- Improving Multimodal Social Media Popularity Prediction via Selective Retrieval Knowledge Augmentation (AAAI 2025 · similarity 0.82)
- SMTPD: A New Benchmark for Temporal Prediction of Social Media Popularity (CVPR 2025 · similarity 0.81)
- Adaptive Elicitation of Latent Information Using Natural Language (ICML 2025 · similarity 0.81)
- ReMoDetect: Reward Models Recognize Aligned LLM's Generations (NeurIPS 2024 · similarity 0.81)
- Public Opinion Field Effect and Hawkes Process Join Hands for Information Popularity Prediction (AAAI 2025 · similarity 0.80)
- Majority of the Bests: Improving Best-of-N via Bootstrapping (NeurIPS 2025 · similarity 0.80)
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.
