Research map / Video / Video generation
Playable Video Generation
Check 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 |
|---|---|---|---|
| Use CADDY's unsupervised discrete action bottleneck and generation framework instead of the strongest reported baseline | BAIR video generation · BAIR | Δ-MSE (%): 80.9 -> 54.8 | Helped |
| Use CADDY instead of the strongest reported baseline | Atari Breakout video generation · Atari Breakout | FVD: 84.4 -> 5.94 | Helped |
| Use CADDY instead of the strongest reported baseline | Tennis video generation · Tennis | LPIPS: 0.104 -> 0.102 | Helped |
| Use CADDY instead of the strongest reported baseline | Tennis video generation · Tennis | FVD: 223 -> 239 | Hurt |
| Add Gumbel-Softmax sampling (ablation) | CADDY · BAIR | Δ-MSE (%): 69.7 -> 64.8 | Helped |
| Add Gumbel-Softmax sampling (ablation) | CADDY · BAIR | Δ-Acc (%): 51.2 -> 37.9 | Hurt |
| Add Gumbel-Softmax sampling and action loss, without action variability embeddings (ablation) | CADDY · BAIR | LPIPS: 0.209 -> 0.249 | Hurt |
| Add Gumbel-Softmax sampling and action loss, without action variability embeddings (ablation) | CADDY · BAIR | Δ-MSE (%): 64.8 -> 92.7 | Hurt |
| Add Gumbel-Softmax sampling and action variability embeddings, without action loss (ablation) | CADDY · BAIR | LPIPS: 0.209 -> 0.245 | Hurt |
| Add Gumbel-Softmax sampling and action variability embeddings, without action loss (ablation) | CADDY · BAIR | Δ-MSE (%): 64.8 -> 93.7 | Hurt |
| Add the complete combination of Gumbel-Softmax sampling, action variability embeddings, and mutual-information action loss (ablation) | CADDY · BAIR | LPIPS: 0.263 -> 0.202 | Helped |
| Add the complete combination of Gumbel-Softmax sampling, action variability embeddings, and mutual-information action loss (ablation) | CADDY · BAIR | Δ-Acc (%): 51.2 -> 69.0 | Helped |
Closest papers by meaning
- GameGen-X: Interactive Open-world Game Video Generation (ICLR 2025 · similarity 0.86)
- Diffusion Models Are Real-Time Game Engines (ICLR 2025 · similarity 0.84)
- Improving Video Generation with Human Feedback (NeurIPS 2025 · similarity 0.83)
- GROOT: Learning to Follow Instructions by Watching Gameplay Videos (ICLR 2024 · similarity 0.83)
- Optical-Flow Guided Prompt Optimization for Coherent Video Generation (CVPR 2025 · similarity 0.83)
- Motion Prompting: Controlling Video Generation with Motion Trajectories (CVPR 2025 · similarity 0.83)
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.