Research map / Video / Video generation
PreLAR: World Model Pre-training with Learnable Action Representation
Lixuan Zhang, Meina Kan, Shiguang Shan, Xilin Chen
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 |
|---|---|---|---|
| pre-train the world model action-conditionally with learnable action representation instead of action-free video prediction | PreLAR vs APV and ContextWM · Meta-world manipulation (RLBench pre-training, aggregate) | success rate vs environment steps | Improved |
| pre-train on real-world SSv2 videos with PreLAR | PreLAR · Meta-world manipulation (SSv2 pre-training, aggregate) | success rate vs environment steps | Improved |
| PreLAR pre-trained on SSv2 and fine-tuned on Door Lock | PreLAR · Meta-world Door Lock (SSv2 pre-training) | success rate | Hurt |
| add action-state consistency loss (ablation) | PreLAR · Meta-world aggregate (RLBench pre-training) | success rate | Improved |
| add action KL loss and stochastic latent action representation (ablation) | PreLAR · Meta-world aggregate (RLBench pre-training) | success rate | Improved |
| naive fine-tune of an action-free pre-trained model instead of action-conditional pre-training (ablation) | world model · Meta-world aggregate (RLBench pre-training) | success rate | No clear effect |
| action-free video pre-training (APV, ContextWM, PreLAR) versus no pre-training | DreamerV2 · Meta-world manipulation, most tasks | success rate and sample efficiency | Improved |
Closest papers by meaning
- AdaWorld: Learning Adaptable World Models with Latent Actions (ICML 2025 · similarity 0.89)
- Zero-shot World Models via Search in Memory (NeurIPS 2025 · similarity 0.88)
- World Action Verifier: Self-Improving World Models via Forward-Inverse Asymmetry (NeurIPS 2026 · similarity 0.88)
- From Forecasting to Planning: Policy World Model for Collaborative State-Action Prediction (NeurIPS 2025 · similarity 0.88)
- Learning an Actionable Discrete Diffusion Policy via Large-Scale Actionless Video Pre-Training (NeurIPS 2024 · similarity 0.87)
- DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning (ICML 2025 · similarity 0.87)
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.
