FloMo: 3D Scene Flow as a World Action Model Intermediate for Learning from Human Video
cs.CV, cs.AI
Submitted: 2026-11-03
Updated: 2026-11-04
Terminology
Sources
- DreamGen: Unlocking Generalization in Robot Learning through Video World Models
- Video Generators are Robot Policies
- World Action Models are Zero-shot Policies
- Any-point Trajectory Modeling for Policy Learning
- AMPLIFY: Actionless Motion Priors for Robot Learning from Videos
- Flow as the Cross-Domain Manipulation Interface
- R3M: A Universal Visual Representation for Robot Manipulation
- VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training
- Robotic Offline RL from Internet Videos via Value-Function Pre-Training
- ViVa: Video-Trained Value Functions for Guiding Online RL from Diverse Data
- Genie: Generative Interactive Environments
- Latent Action Pretraining from Videos
- Moto: Latent Motion Token as the Bridging Language for Learning Robot Manipulation from Videos
- EgoBridge: Domain Adaptation for Generalizable Imitation from Egocentric Human Data
- EgoMimic: Scaling Imitation Learning via Egocentric Video
- UniSkill: Imitating Human Videos via Cross-Embodiment Skill Representations
- OKAMI: Teaching Humanoid Robots Manipulation Skills through Single Video Imitation
- X-Sim: Cross-Embodiment Learning via Real-to-Sim-to-Real
- General Flow as Foundation Affordance for Scalable Robot Learning
- FlowRetrieval: Flow-Guided Data Retrieval for Few-Shot Imitation Learning
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs