IronMind: Scaling Humanoid Dexterous Manipulation via Camera-Space Ego-Centric Pretraining
cs.RO
Submitted: 2026-09-30
Updated: 2026-09-30
Comments: https://xpeng-robotics.github.io/ironmind/
Project page: https://xpeng-robotics.github.io/ironmind/Abstract
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- AgiBot World Colosseo: A Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systems
- Qwen2.5-VL Technical Report
- Data Quality in Imitation Learning
- $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control
- RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control
- InternVLA-A1: Unifying Understanding, Generation and Action for Robotic Manipulation
- Translation as a Bridging Action: Transferring Manipulation Skills from Humans to Robots
- DIAL: Decoupling Intent and Action via Latent World Modeling for End-to-End VLA
- Capek 0.5: An Execution-Centric Vision-Language Model for Embodied Intelligence
- Diffusion Policy: Visuomotor Policy Learning via Action Diffusion
- Rescaling Egocentric Vision
- MHR: Momentum Human Rig
- VLA-0: Building State-of-the-Art VLAs with Zero Modification
- Ego4D: Around the World in 3,000 Hours of Egocentric Video
- Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectives
- Worlds in One Demo: A Synthetic Data Engine for Learning Open-World Mobile Manipulation
- Re-Mix: Optimizing Data Mixtures for Large Scale Imitation Learning
- Omni-View: Unlocking How Generation Facilitates Understanding in Unified 3D Model based on Multiview images
- Galaxea Open-World Dataset and G0 Dual-System VLA Model
- Scaling Laws for Neural Language Models
Related papers
- FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
- MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
- RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
- ARC: A Reasoning Recipe for Robot Foundation Models
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies