# Video2Policy: Scaling up Manipulation Tasks in Simulation through Internet Videos

arXiv: https://arxiv.org/abs/2502.09886 · HTML: https://mlsanity.com/paper/2502.09886/

## Key facts

- Cited by 26 papers in the map (data through 14 Oct 2026).
- Placed in Human-video and latent-action pretraining, one of 13 directions in Vision-language-action models.
- Builds on 44 of its 51 references that are in the map.
- First posted on arXiv 14 Feb 2025.

## Placed in

- Vision-language-action models / [Human-video and latent-action pretraining](https://mlsanity.com/vla/human-video-and-latent-action-pretraining/)

## Builds on (most cited first, 30 of 44)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [DreamFusion: Text-to-3D using 2D Diffusion](https://mlsanity.com/paper/2209.14988/) (2022)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [Do As I Can, Not As I Say: Grounding Language in Robotic Affordances](https://mlsanity.com/paper/2204.01691/) (2022)
- [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://mlsanity.com/paper/2305.16291/) (2023)
- [Code Llama: Open Foundation Models for Code](https://mlsanity.com/paper/2308.12950/) (2023)
- [Octo: An Open-Source Generalist Robot Policy](https://mlsanity.com/paper/2405.12213/) (2024)
- [Mastering Diverse Domains through World Models](https://mlsanity.com/paper/2301.04104/) (2023)
- [Code as Policies: Language Model Programs for Embodied Control](https://mlsanity.com/paper/2209.07753/) (2022)
- [The "something something" video database for learning and evaluating visual common sense](https://mlsanity.com/paper/1706.04261/) (2017)
- [Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning](https://mlsanity.com/paper/2108.10470/) (2021)
- [Open X-Embodiment: Robotic Learning Datasets and RT-X Models](https://mlsanity.com/paper/2310.08864/) (2023)
- [A Generalist Agent](https://mlsanity.com/paper/2205.06175/) (2022)
- [VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models](https://mlsanity.com/paper/2307.05973/) (2023)
- [Learning Dexterous In-Hand Manipulation](https://mlsanity.com/paper/1808.00177/) (2018)
- [Eureka: Human-Level Reward Design via Coding Large Language Models](https://mlsanity.com/paper/2310.12931/) (2023)
- [Learning Universal Policies via Text-Guided Video Generation](https://mlsanity.com/paper/2302.00111/) (2023)
- [Learning agile and dynamic motor skills for legged robots](https://mlsanity.com/paper/1901.08652/) (2019)
- [InstantMesh: Efficient 3D Mesh Generation from a Single Image with Sparse-view Large Reconstruction Models](https://mlsanity.com/paper/2404.07191/) (2024)
- [FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects](https://mlsanity.com/paper/2312.08344/) (2023)
- [RoboCasa: Large-Scale Simulation of Everyday Tasks for Generalist Robots](https://mlsanity.com/paper/2406.02523/) (2024)
- [UniDepth: Universal Monocular Metric Depth Estimation](https://mlsanity.com/paper/2403.18913/) (2024)
- [ProcTHOR: Large-Scale Embodied AI Using Procedural Generation](https://mlsanity.com/paper/2206.06994/) (2022)
- [ManiSkill2: A Unified Benchmark for Generalizable Manipulation Skills](https://mlsanity.com/paper/2302.04659/) (2023)
- [Mastering Atari Games with Limited Data](https://mlsanity.com/paper/2111.00210/) (2021)
- [Text2Motion: From Natural Language Instructions to Feasible Plans](https://mlsanity.com/paper/2303.12153/) (2023)

## Built on by (most cited first, 26 of 26)

- [Motus: A Unified Latent Action World Model](https://mlsanity.com/paper/2512.13030/) (2025)
- [X-Sim: Cross-Embodiment Learning via Real-to-Sim-to-Real](https://mlsanity.com/paper/2505.07096/) (2025)
- [ObjectVLA: End-to-End Open-World Object Manipulation Without Demonstration](https://mlsanity.com/paper/2502.19250/) (2025)
- [Robot Learning from a Physical World Model](https://mlsanity.com/paper/2511.07416/) (2025)
- [ActiveUMI: Robotic Manipulation with Active Perception from Robot-Free Human Demonstrations](https://mlsanity.com/paper/2510.01607/) (2025)
- [Dexterous Manipulation Policies from RGB Human Videos via 3D Hand-Object Trajectory Reconstruction](https://mlsanity.com/paper/2602.09013/) (2026)
- [OXE-AugE: A Large-Scale Robot Augmentation of OXE for Scaling Cross-Embodiment Policy Learning](https://mlsanity.com/paper/2512.13100/) (2025)
- [Vision-Language-Action in Robotics: A Survey of Datasets, Benchmarks, and Data Engines](https://mlsanity.com/paper/2604.23001/) (2026)
- [Robot Learning from Any Images](https://mlsanity.com/paper/2509.22970/) (2025)
- [ViSA-Flow: Accelerating Robot Skill Learning via Large-Scale Video Semantic Action Flow](https://mlsanity.com/paper/2505.01288/) (2025)
- [TwinRL: Digital Twin-Driven Reinforcement Learning for Real-World Robotic Manipulation](https://mlsanity.com/paper/2602.09023/) (2026)
- [SimLauncher: Launching Sample-Efficient Real-world Robotic Reinforcement Learning via Simulation Pre-training](https://mlsanity.com/paper/2507.04452/) (2025)
- [Robot Learning from Human Videos: A Survey](https://mlsanity.com/paper/2604.27621/) (2026)
- [LodeStar: Long-horizon Dexterity via Synthetic Data Augmentation from Human Demonstrations](https://mlsanity.com/paper/2508.17547/) (2025)
- [How Far Are Vision-Language Models from Constructing the Real World? A Benchmark for Physical Generative Reasoning](https://mlsanity.com/paper/2603.24866/) (2026)
- [IGen: Scalable Data Generation for Robot Learning from Open-World Images](https://mlsanity.com/paper/2512.01773/) (2025)
- [RoboGene: Boosting VLA Pre-training via Diversity-Driven Agentic Framework for Real-World Task Generation](https://mlsanity.com/paper/2602.16444/) (2026)
- [KnowDemo: Knowledge-Guided Robot Demonstration Generation from Human Videos](https://mlsanity.com/paper/2609.21229/) (2026)
- [Cross-Domain Demo-to-Code via Neurosymbolic Counterfactual Reasoning](https://mlsanity.com/paper/2603.18495/) (2026)
- [CoRe: Combined Rewards with Vision-Language Model Feedback for Preference-Aligned Reinforcement Learning](https://mlsanity.com/paper/2607.01721/) (2026)
- [Test-Time Spatial Reasoning for Robot Manipulation Using Generative Real-to-Sim](https://mlsanity.com/paper/2609.33982/) (2026)
- [ZeroBot: Learning from Scratch in Minutes with Generative Real2Sim](https://mlsanity.com/paper/2609.34010/) (2026)
- [Dex-One2Many: Learning Dexterous Manipulation from a Single Human Demonstration](https://mlsanity.com/paper/2610.12470/) (2026)
- [DexCast: Learning Real-World Dexterous Control from Human Videos](https://mlsanity.com/paper/2610.14691/) (2026)
- [RoboSimStudio: An Interactive Toolkit for Human and Agent Rapid Real2Sim of Robot Manipulation Tasks](https://mlsanity.com/paper/2610.18835/) (2026)
- [Human Demonstrations as Post-Training Data for Any Robot Body: Retargeting the Object, Not the Hand](https://mlsanity.com/paper/2610.20351/) (2026)

## Abstract (arXiv)

Simulation offers a promising approach for cheaply scaling training data for generalist policies. To scalably generate data from diverse and realistic tasks, existing algorithms either rely on large language models (LLMs) that may hallucinate tasks not interesting for robotics; or digital twins, which require careful real-to-sim alignment and are hard to scale. To address these challenges, we introduce Video2Policy, a novel framework that leverages internet RGB videos to reconstruct tasks based on everyday human behavior. Our approach comprises two phases: (1) task generation in simulation from videos; and (2) reinforcement learning utilizing in-context LLM-generated reward functions iteratively. We demonstrate the efficacy of Video2Policy by reconstructing over 100 videos from the Something-Something-v2 (SSv2) dataset, which depicts diverse and complex human behaviors on 9 different tasks. Our method can successfully train RL policies on such tasks, including complex and challenging tasks such as throwing. Finally, we show that the generated simulation data can be scaled up for training a general policy, and it can be transferred back to the real robot in a Real2Sim2Real way.