Research map / Large language models / Retrieval and knowledge
Story Ending Generation with Incremental Encoding and Commonsense Knowledge
Jian Guan, Yansen Wang, Minlie Huang
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 incremental encoding instead of concatenated or hierarchical context encoding | Story ending generation on ROCStories · ROCStories test, manual sample | Logicality: 1.10 vs. 0.70/0.84 | Improved |
| Use incremental encoding instead of hierarchical encoding | HLSTM+MSA · ROCStories test, manual sample | Logicality: 1.26 vs. 1.06; 1.24 vs. 1.02 | Improved |
| Add commonsense knowledge context vectors using graph attention (ablation) | IE · ROCStories test, manual sample | Logicality: 1.10 vs. 1.26 | Improved |
| Add commonsense knowledge context vectors using contextual attention (ablation) | IE · ROCStories test, manual sample | Logicality: 1.10 vs. 1.24 | Improved |
| Add commonsense knowledge attention | HLSTM · ROCStories test, manual sample | Logicality: 0.84 vs. 1.06; 0.84 vs. 1.02 | Improved |
| Add commonsense knowledge attention | HLSTM+Copy · ROCStories test, manual sample | Logicality: 0.90 vs. 1.06; 0.90 vs. 1.02 | Improved |
| Replace graph attention with contextual attention for graph representation (ablation) | IE+MSA · ROCStories test, manual sample | Logicality: 1.26 vs. 1.24 | No clear effect |
| Replace graph attention with contextual attention for graph representation (ablation) | IE+MSA · ROCStories test | BLEU-1: 0.2566 -> 0.2682 | Improved |
| Add incremental encoding to the encoder | Story ending generation on ROCStories · ROCStories test | PPL: 18.97 -> 11.04 | Improved |
| Add incremental encoding to the encoder | Story ending generation on ROCStories · ROCStories test, manual sample | Grammar: 1.74 -> 1.84; 1.57 -> 1.84 | Improved |
| Add incremental-encoding supervision to the encoder | Story ending generation model · ROCStories | Logicality | Improved |
| Add a copy mechanism | HLSTM · ROCStories test | PPL: 17.26 -> 19.93 | Hurt |
Closest papers by meaning
- Incorporating Structured Commonsense Knowledge in Story Completion (AAAI 2019 · similarity 0.90)
- Automated Storytelling via Causal, Commonsense Plot Ordering (AAAI 2021 · similarity 0.85)
- Discriminative Sentence Modeling for Story Ending Prediction (AAAI 2020 · similarity 0.84)
- StoryWeaver: A Unified World Model for Knowledge-Enhanced Story Character Customization (AAAI 2025 · similarity 0.82)
- Agents' Room: Narrative Generation through Multi-step Collaboration (ICLR 2025 · similarity 0.82)
- Intelligent Grimm - Open-ended Visual Storytelling via Latent Diffusion Models (CVPR 2024 · similarity 0.81)
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.
