# Learning Object Manipulation Skills via Approximate State Estimation from Real Videos

arXiv: https://arxiv.org/abs/2011.06813 · HTML: https://mlsanity.com/paper/2011.06813/

## Key facts

- Cited by 12 papers in the map (data through 7 Oct 2026).
- First posted on arXiv 13 Nov 2020.

## Built on by (most cited first, 12 of 12)

- [Learning Generalizable Robotic Reward Functions from "In-The-Wild" Human Videos](https://mlsanity.com/paper/2103.16817/) (2021)
- [Instruction-driven history-aware policies for robotic manipulations](https://mlsanity.com/paper/2209.04899/) (2022)
- [Learning by Watching: Physical Imitation of Manipulation Skills from Human Videos](https://mlsanity.com/paper/2101.07241/) (2021)
- [Human Hands as Probes for Interactive Object Understanding](https://mlsanity.com/paper/2112.09120/) (2021)
- [Vid2Robot: End-to-end Video-conditioned Policy Learning with Cross-Attention Transformers](https://mlsanity.com/paper/2403.12943/) (2024)
- [Learning to Imitate Object Interactions from Internet Videos](https://mlsanity.com/paper/2211.13225/) (2022)
- [Learning Video-Conditioned Policies for Unseen Manipulation Tasks](https://mlsanity.com/paper/2305.06289/) (2023)
- [Learning Reward Functions for Robotic Manipulation by Observing Humans](https://mlsanity.com/paper/2211.09019/) (2022)
- [Learning by Watching: A Review of Video-based Learning Approaches for Robot Manipulation](https://mlsanity.com/paper/2402.07127/) (2024)
- [6D Object Pose Tracking in Internet Videos for Robotic Manipulation](https://mlsanity.com/paper/2503.10307/) (2025)
- [Learning Object Manipulation Skills from Video via Approximate Differentiable Physics](https://mlsanity.com/paper/2208.01960/) (2022)
- [Back to Reality for Imitation Learning](https://mlsanity.com/paper/2111.12867/) (2021)

## Abstract (arXiv)

Humans are adept at learning new tasks by watching a few instructional videos. On the other hand, robots that learn new actions either require a lot of effort through trial and error, or use expert demonstrations that are challenging to obtain. In this paper, we explore a method that facilitates learning object manipulation skills directly from videos. Leveraging recent advances in 2D visual recognition and differentiable rendering, we develop an optimization based method to estimate a coarse 3D state representation for the hand and the manipulated object(s) without requiring any supervision. We use these trajectories as dense rewards for an agent that learns to mimic them through reinforcement learning. We evaluate our method on simple single- and two-object actions from the Something-Something dataset. Our approach allows an agent to learn actions from single videos, while watching multiple demonstrations makes the policy more robust. We show that policies learned in a simulated environment can be easily transferred to a real robot.