# DexPolicy: Scheduled Exploration for Trajectory-Guided Dexterous Manipulation

arXiv: https://arxiv.org/abs/2610.00360 · HTML: https://mlsanity.com/paper/2610.00360/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 22 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Builds on (most cited first, 19 of 19)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Trust Region Policy Optimization](https://mlsanity.com/paper/1502.05477/) (2015)
- [High-Dimensional Continuous Control Using Generalized Advantage Estimation](https://mlsanity.com/paper/1506.02438/) (2015)
- [Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations](https://mlsanity.com/paper/1709.10087/) (2017)
- [Learning Dexterous In-Hand Manipulation](https://mlsanity.com/paper/1808.00177/) (2018)
- [FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects](https://mlsanity.com/paper/2312.08344/) (2023)
- [DexMV: Imitation Learning for Dexterous Manipulation from Human Videos](https://mlsanity.com/paper/2108.05877/) (2021)
- [Flow Matching Policy Gradients](https://mlsanity.com/paper/2507.21053/) (2025)
- [ReinFlow: Fine-tuning Flow Matching Policy with Online Reinforcement Learning](https://mlsanity.com/paper/2505.22094/) (2025)
- [DexVIP: Learning Dexterous Grasping with Human Hand Pose Priors from Video](https://mlsanity.com/paper/2202.00164/) (2022)
- [Smooth Exploration for Robotic Reinforcement Learning](https://mlsanity.com/paper/2005.05719/) (2020)
- [ViViDex: Learning Vision-based Dexterous Manipulation from Human Videos](https://mlsanity.com/paper/2404.15709/) (2024)
- [Action Noise in Off-Policy Deep Reinforcement Learning: Impact on Exploration and Performance](https://mlsanity.com/paper/2206.03787/) (2022)
- [DexTrack: Towards Generalizable Neural Tracking Control for Dexterous Manipulation from Human References](https://mlsanity.com/paper/2502.09614/) (2025)
- [Flow Policy Gradients for Robot Control](https://mlsanity.com/paper/2602.02481/) (2026)
- [PPO-CMA: Proximal Policy Optimization with Covariance Matrix Adaptation](https://mlsanity.com/paper/1810.02541/) (2018)
- [Colored Noise in PPO: Improved Exploration and Performance through Correlated Action Sampling](https://mlsanity.com/paper/2312.11091/) (2023)

## Abstract (arXiv)

Reinforcement learning (RL) for dexterous manipulation must discover finger-object contacts and then control the object precisely; the action noise that serves the first goal can interfere with the second. In trajectory-guided settings such as ViViDex, where RL refine hand-object trajectories from human video, our baseline PPO runs end near their initial action noise after 5M steps, motivating explicit control of exploration scale. DexPolicy makes that scale an explicit function of training steps, annealing from broad to narrow exploration while holding loss, architecture, reward, and optimizer settings fixed. We study three policy-optimization settings: PPO, critic-free GRPO continuation, and a flow-parameterized PPO variant (FPO). Across five YCB objects and three training seeds, mean deterministic Target success rises from 49.4% to 68.1% (FPO), 14.1% to 45.4% (GRPO), and 32.0% to 35.7% (PPO). On a RealMan RM75 arm with an Inspire/RH56 hand, 360 trials over three objects raise mean Target success from 25.0% to 85.0% (FPO), 10.0% to 63.3% (GRPO), and 8.3% to 43.3% (PPO), with one trained model per object-method condition. PPO component screening favors noise control over the tested optimizer contraction; the selected PPO schedule yields higher mean Target success than linear decay with the same endpoints on three tested objects. Training return, deterministic Target success, and tolerance to execution noise dissociate; schedules should therefore be judged by terminal task success under the intended execution conditions, per task and policy-optimization setting. Code: https://github.com/AIGeeksGroup/DexPolicy. Website: https://aigeeksgroup.github.io/DexPolicy.