# DexHiL: A Human-in-the-Loop Framework for Vision-Language-Action Model Post-Training in Dexterous Manipulation

arXiv: https://arxiv.org/abs/2603.09121 · HTML: https://mlsanity.com/paper/2603.09121/

## Key facts

- Cited by 6 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 33 of its 35 references that are in the map.
- First posted on arXiv 10 Mar 2026.

## 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 33)

- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [RT-1: Robotics Transformer for Real-World Control at Scale](https://mlsanity.com/paper/2212.06817/) (2022)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [Octo: An Open-Source Generalist Robot Policy](https://mlsanity.com/paper/2405.12213/) (2024)
- [Vision-Language Foundation Models as Effective Robot Imitators](https://mlsanity.com/paper/2311.01378/) (2023)
- [HG-DAgger: Interactive Imitation Learning with Human Experts](https://mlsanity.com/paper/1810.02890/) (2018)
- [GELLO: A General, Low-Cost, and Intuitive Teleoperation Framework for Robot Manipulators](https://mlsanity.com/paper/2309.13037/) (2023)
- [DexCap: Scalable and Portable Mocap Data Collection System for Dexterous Manipulation](https://mlsanity.com/paper/2403.07788/) (2024)
- [ConRFT: A Reinforced Fine-tuning Method for VLA Models via Consistency Policy](https://mlsanity.com/paper/2502.05450/) (2025)
- [AnyTeleop: A General Vision-Based Dexterous Robot Arm-Hand Teleoperation System](https://mlsanity.com/paper/2307.04577/) (2023)
- [DexPilot: Vision Based Teleoperation of Dexterous Robotic Hand-Arm System](https://mlsanity.com/paper/1910.03135/) (2019)
- [ThriftyDAgger: Budget-Aware Novelty and Risk Gating for Interactive Imitation Learning](https://mlsanity.com/paper/2109.08273/) (2021)
- [Robot Learning on the Job: Human-in-the-Loop Autonomy and Learning During Deployment](https://mlsanity.com/paper/2211.08416/) (2022)
- [EnsembleDAgger: A Bayesian Approach to Safe Imitation Learning](https://mlsanity.com/paper/1807.08364/) (2018)
- [DexGraspVLA: A Vision-Language-Action Framework Towards General Dexterous Grasping](https://mlsanity.com/paper/2502.20900/) (2025)
- [HOMIE: Humanoid Loco-Manipulation with Isomorphic Exoskeleton Cockpit](https://mlsanity.com/paper/2502.13013/) (2025)
- [Scalable Vision-Language-Action Model Pretraining for Robotic Manipulation with Real-Life Human Activity Videos](https://mlsanity.com/paper/2510.21571/) (2025)
- [ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning](https://mlsanity.com/paper/2503.21860/) (2025)
- [Being-H0.5: Scaling Human-Centric Robot Learning for Cross-Embodiment Generalization](https://mlsanity.com/paper/2601.12993/) (2026)
- [DexMachina: Functional Retargeting for Bimanual Dexterous Manipulation](https://mlsanity.com/paper/2505.24853/) (2025)
- [Diffusion Meets DAgger: Supercharging Eye-in-hand Imitation Learning](https://mlsanity.com/paper/2402.17768/) (2024)
- [ACE: A Cross-Platform Visual-Exoskeletons System for Low-Cost Dexterous Teleoperation](https://mlsanity.com/paper/2408.11805/) (2024)
- [SPIDER: Scalable Physics-Informed Dexterous Retargeting](https://mlsanity.com/paper/2511.09484/) (2025)
- [DexVLG: Dexterous Vision-Language-Grasp Model at Scale](https://mlsanity.com/paper/2507.02747/) (2025)
- [MEGA-DAgger: Imitation Learning with Multiple Imperfect Experts](https://mlsanity.com/paper/2303.00638/) (2023)
- [Geometric Retargeting: A Principled, Ultrafast Neural Hand Retargeting Algorithm](https://mlsanity.com/paper/2503.07541/) (2025)
- [Dexplore: Scalable Neural Control for Dexterous Manipulation from Reference-Scoped Exploration](https://mlsanity.com/paper/2509.09671/) (2025)
- [Dexterous Teleoperation of 20-DoF ByteDexter Hand via Human Motion Retargeting](https://mlsanity.com/paper/2507.03227/) (2025)

## Built on by (most cited first, 6 of 6)

- [Orca: The World is in Your Mind](https://mlsanity.com/paper/2606.30534/) (2026)
- [BORA: Bridging Offline Reinforcement Learning and Online Residual Adaptation for Real-World Dexterous VLA Models](https://mlsanity.com/paper/2605.30226/) (2026)
- [Unified Video-Action Joint Denoising for Dexterous Action and Data Generation](https://mlsanity.com/paper/2606.03868/) (2026)
- [DexPIE: Stable Dexterous Policy Improvement from Real-World Experience](https://mlsanity.com/paper/2606.09615/) (2026)
- [ROVE: Unlocking Human Interventions for Humanoid Manipulation via Reinforcement Learning](https://mlsanity.com/paper/2606.17011/) (2026)
- [I3L: Iterative Interactive Imitation Learning for Whole-Body Mobile Manipulation](https://mlsanity.com/paper/2610.17748/) (2026)

## Abstract (arXiv)

While Vision-Language-Action (VLA) models have demonstrated promising generalization capabilities in robotic manipulation, deploying them on specific and complex downstream tasks still demands effective post-training. In parallel, Human-in-the-Loop (HiL) learning has proven to be a powerful mechanism for refining robot policies. However, extending this paradigm to dexterous manipulation remains challenging: multi-finger control is high-dimensional, contact-intensive, and exhibits execution distributions that differ markedly from standard arm motions, leaving existing dexterous VLA systems limited in reliability and adaptability. We present DexHiL, the first integrated arm-hand human-in-the-loop framework for dexterous VLA models, enabling coordinated interventions over the arm and the dexterous hand within a single system. DexHiL introduces an intervention-aware data sampling strategy that prioritizes corrective segments for post-training, alongside a lightweight teleoperation interface that supports instantaneous human corrections during execution. Real-robot experiments demonstrate that DexHiL serves as an effective post-training framework, yielding a substantial performance leap, outperforming standard offline-only fine-tuning baselines by an average of 25% in success rates across distinct tasks. Project page: https://chenzhongxi-sjtu.github.io/dexhil/