# End-to-end RL Improves Dexterous Grasping Policies

arXiv: https://arxiv.org/abs/2509.16434 · HTML: https://mlsanity.com/paper/2509.16434/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 28 references that are in the map.
- First posted on arXiv 19 Sep 2025.

## Builds on (most cited first, 26 of 26)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning](https://mlsanity.com/paper/2108.10470/) (2021)
- [IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures](https://mlsanity.com/paper/1802.01561/) (2018)
- [End-to-End Training of Deep Visuomotor Policies](https://mlsanity.com/paper/1504.00702/) (2015)
- [QT-Opt: Scalable Deep Reinforcement Learning for Vision-Based Robotic Manipulation](https://mlsanity.com/paper/1806.10293/) (2018)
- [Solving Rubik's Cube with a Robot Hand](https://mlsanity.com/paper/1910.07113/) (2019)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [Learning Quadrupedal Locomotion over Challenging Terrain](https://mlsanity.com/paper/2010.11251/) (2020)
- [RMA: Rapid Motor Adaptation for Legged Robots](https://mlsanity.com/paper/2107.04034/) (2021)
- [Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots](https://mlsanity.com/paper/2402.10329/) (2024)
- [Learning robust perceptive locomotion for quadrupedal robots in the wild](https://mlsanity.com/paper/2201.08117/) (2022)
- [Orbit: A Unified Simulation Framework for Interactive Robot Learning Environments](https://mlsanity.com/paper/2301.04195/) (2023)
- [ManiSkill3: GPU Parallelized Robotics Simulation and Rendering for Generalizable Embodied AI](https://mlsanity.com/paper/2410.00425/) (2024)
- [Sim-to-Real via Sim-to-Sim: Data-efficient Robotic Grasping via Randomized-to-Canonical Adaptation Networks](https://mlsanity.com/paper/1812.07252/) (2018)
- [Legged Locomotion in Challenging Terrains using Egocentric Vision](https://mlsanity.com/paper/2211.07638/) (2022)
- [Sim-to-Real Reinforcement Learning for Deformable Object Manipulation](https://mlsanity.com/paper/1806.07851/) (2018)
- [Visual Dexterity: In-Hand Reorientation of Novel and Complex Object Shapes](https://mlsanity.com/paper/2211.11744/) (2022)
- [MuJoCo Playground](https://mlsanity.com/paper/2502.08844/) (2025)
- [Learning Visuotactile Skills with Two Multifingered Hands](https://mlsanity.com/paper/2404.16823/) (2024)
- [Transferring End-to-End Visuomotor Control from Simulation to Real World for a Multi-Stage Task](https://mlsanity.com/paper/1707.02267/) (2017)
- [Sim-to-Real Reinforcement Learning for Vision-Based Dexterous Manipulation on Humanoids](https://mlsanity.com/paper/2502.20396/) (2025)
- [DextrAH-RGB: Visuomotor Policies to Grasp Anything with Dexterous Hands](https://mlsanity.com/paper/2412.01791/) (2024)
- [DextrAH-G: Pixels-to-Action Dexterous Arm-Hand Grasping with Geometric Fabrics](https://mlsanity.com/paper/2407.02274/) (2024)
- [RobustDexGrasp: Robust Dexterous Grasping of General Objects](https://mlsanity.com/paper/2504.05287/) (2025)
- [Distilling Realizable Students from Unrealizable Teachers](https://mlsanity.com/paper/2505.09546/) (2025)

## Built on by (most cited first, 5 of 5)

- [SimToolReal: An Object-Centric Policy for Zero-Shot Dexterous Tool Manipulation](https://mlsanity.com/paper/2602.16863/) (2026)
- [Dex4D: Task-Agnostic Point Track Policy for Zero-Shot Dexterous Manipulation](https://mlsanity.com/paper/2602.15828/) (2026)
- [DeFM: Learning Foundation Representations from Depth for Robotics](https://mlsanity.com/paper/2601.18923/) (2026)
- [Efficient On-policy Visual-RL via Stochastic Decoupled Policy Gradient](https://mlsanity.com/paper/2605.26478/) (2026)
- [Play2Perfect: What Matters in Dexterous Play Pretraining for Precise Assembly?](https://mlsanity.com/paper/2606.26428/) (2026)

## Abstract (arXiv)

This work explores techniques to scale up image-based end-to-end learning for dexterous grasping with an arm + hand system. Unlike state-based RL, vision-based RL is much more memory inefficient, resulting in relatively low batch sizes, which is not amenable for algorithms like PPO. Nevertheless, it is still an attractive method as unlike the more commonly used techniques which distill state-based policies into vision networks, end-to-end RL can allow for emergent active vision behaviors. We identify a key bottleneck in training these policies is the way most existing simulators scale to multiple GPUs using traditional data parallelism techniques. We propose a new method where we disaggregate the simulator and RL (both training and experience buffers) onto separate GPUs. On a node with four GPUs, we have the simulator running on three of them, and PPO running on the fourth. We are able to show that with the same number of GPUs, we can double the number of existing environments compared to the previous baseline of standard data parallelism. This allows us to train vision-based environments, end-to-end with depth, which were previously performing far worse with the baseline. We train and distill both depth and state-based policies into stereo RGB networks and show that depth distillation leads to better results, both in simulation and reality. This improvement is likely due to the observability gap between state and vision policies which does not exist when distilling depth policies into stereo RGB. We further show that the increased batch size brought about by disaggregated simulation also improves real world performance. When deploying in the real world, we improve upon the previous state-of-the-art vision-based results using our end-to-end policies.