Research map / Large language models / Agents, code and math
On the Wings of Imagination: Conflicting Script-based Multi-role Framework for Humor Caption Generation
Wenbo Shang, Yuxi Sun, Jing Ma, Xin 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 |
|---|---|---|---|
| add conflict scripts to I+D+T_im (ablation) | HOMER (GPT-4o) · Humor in AI #Top10 | pass@1 (%): 34.40 -> 66.41 | Improved |
| add imagination tree to I+D+C (ablation) | HOMER (GPT-4o) · Humor in AI #Top10 | pass@1 (%): 57.40 -> 66.41 | Improved |
| add situation description to I+C+T_im (ablation) | HOMER (GPT-4o) · Humor in AI #Top10 | pass@1 (%): 35.40 -> 66.41 | Improved |
| add imagination tree alone to image-only prompt (ablation) | Image-only generator (GPT-4o) · Humor in AI #Top10 | pass@1 (%): 20.20 -> 20.00 | No clear effect |
| add situation description and conflict scripts to image-only (ablation) | Generator (GPT-4o) · Humor in AI #Top10 | pass@1 (%): 20.20 -> 57.40 | Improved |
| add imagination tree to I+D+C (ablation) | HOMER (GPT-4o) · Electric sheep Better Group | pass@1 (%): 60.33 -> 75.53 | Improved |
| remove relevance, frequency or diversity score from humor-relevance pruning (ablation) | HOMER · Humor in AI / Electric sheep | pass@1 (%) | Hurt |
| vary retrieved jokes k and entities delta over 3,5,7,9 (ablation) | HOMER · Humor in AI / Electric sheep | pass@1 (%) | No clear effect |
| HOMER vs best baselines with GPT-4o | HOMER · Humor in AI, Electric sheep | pass@1 (%): +8.62% pass@1 average | Improved |
| HOMER vs baselines on ImgFlip memes | HOMER · ImgFlip meme | pass@1 (%): 76.67 -> 83.33 | Improved |
Closest papers by meaning
- HUMORCHAIN: Theory-Guided Multi-Stage Reasoning for Interpretable Multimodal Humor Generation (CVPR 2026 · similarity 0.92)
- From Perception to Punchline: Empowering VLM with the Art of In-the-wild Memes (NeurIPS 2026 · similarity 0.90)
- Humor in AI: Massive Scale Crowd-Sourced Preferences and Benchmarks for Cartoon Captioning (NeurIPS 2024 · similarity 0.89)
- Innovative Thinking, Infinite Humor: Humor Research of Large Language Models through Structured Thought Leaps (ICLR 2025 · similarity 0.87)
- Cracking the Code of Juxtaposition: Can AI Models Understand the Humorous Contradictions (NeurIPS 2024 · similarity 0.87)
- Talk Funny! A Large-Scale Humor Response Dataset with Chain-of-Humor Interpretation (AAAI 2024 · similarity 0.85)
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.
