Masking the end-effector from wrist views during training lets a single-gripper VLA transfer zero-shot to other grippers, arms, and five-fingered hands while keeping original performance.
Scaling cross- embodied learning: One policy for manipulation, navigation, locomotion and aviation
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
abstract
Modern machine learning systems rely on large datasets to attain broad generalization, and this often poses a challenge in robot learning, where each robotic platform and task might have only a small dataset. By training a single policy across many different kinds of robots, a robot learning method can leverage much broader and more diverse datasets, which in turn can lead to better generalization and robustness. However, training a single policy on multi-robot data is challenging because robots can have widely varying sensors, actuators, and control frequencies. We propose CrossFormer, a scalable and flexible transformer-based policy that can consume data from any embodiment. We train CrossFormer on the largest and most diverse dataset to date, 900K trajectories across 20 different robot embodiments. We demonstrate that the same network weights can control vastly different robots, including single and dual arm manipulation systems, wheeled robots, quadcopters, and quadrupeds. Unlike prior work, our model does not require manual alignment of the observation or action spaces. Extensive experiments in the real world show that our method matches the performance of specialist policies tailored for each embodiment, while also significantly outperforming the prior state of the art in cross-embodiment learning.
citation-role summary
method 1
citation-polarity summary
roles
method 1polarities
use method 1representative citing papers
Demo-JEPA enables one-shot cross-embodiment imitation by mapping visual demonstrations to shared latent future trajectories that serve as subgoals for the target agent's own forward dynamics planning.
A consortium released the largest open medical robotics dataset spanning 50+ institutions and used it to train an open VLA model achieving 25% full suturing completion and a multi-embodiment surgical world model.
A modular benchmark of 100 dexterous manipulation tasks across 3 arms and 6 hands with 3,180 demonstrations reveals that current policies (Diffusion Policy, DP3, OpenVLA, π0.5) achieve only 34% mean success, exposing unsolved challenges in contact-rich and precise manipulation.
InDex adapts parallel-gripper VLA policies to dexterous hands by conditioning a diffusion hand policy on a normalized scalar grasp intent, raising average simulated task success from 50.3% to 85.8%.
RECENT decouples skill semantics from embodiment-specific bindings via code refactoring to let small language models achieve skill grounding performance matching large language model baselines.
MonoDuo generates synthetic bimanual demonstrations from single-arm teleoperation plus human collaboration to train policies achieving up to 70% zero-shot success on five manipulation tasks, with 65-70% gains from 25-shot finetuning.
Sim-and-real co-training for robot policies is driven primarily by balanced cross-domain representation alignment and secondarily by domain-dependent action reweighting.
mimic-video combines internet video pretraining with a flow-matching decoder to achieve state-of-the-art robotic manipulation performance with 10x better sample efficiency than vision-language-action models.
DreamPolicy integrates an autoregressive diffusion world model with policy learning to produce a single scalable policy that generalizes to unseen composite terrains for humanoid locomotion.
KITE decouples task reasoning from embodiment-specific control via learned latent interaction intents to enable zero-shot transfer across structurally different robots.
GR-3 is a VLA model that generalizes to novel objects, environments, and abstract instructions, outperforms the π0 baseline, and integrates with the new ByteMini bi-manual mobile robot.
citing papers explorer
-
Cloak: Zero-Shot Cross-Embodiment Manipulation by Masking the End-Effector from the VLA
Masking the end-effector from wrist views during training lets a single-gripper VLA transfer zero-shot to other grippers, arms, and five-fingered hands while keeping original performance.
-
Demo-JEPA: Joint-Embedding Predictive Architecture for One-shot Cross-Embodiment Imitation
Demo-JEPA enables one-shot cross-embodiment imitation by mapping visual demonstrations to shared latent future trajectories that serve as subgoals for the target agent's own forward dynamics planning.
-
Open-H-Embodiment: A Large-Scale Dataset for Enabling Foundation Models in Medical Robotics
A consortium released the largest open medical robotics dataset spanning 50+ institutions and used it to train an open VLA model achieving 25% full suturing completion and a multi-embodiment surgical world model.
-
DexVerse: A Modular Benchmark for Multi-Task, Multi-Embodiment Dexterous Manipulation
A modular benchmark of 100 dexterous manipulation tasks across 3 arms and 6 hands with 3,180 demonstrations reveals that current policies (Diffusion Policy, DP3, OpenVLA, π0.5) achieve only 34% mean success, exposing unsolved challenges in contact-rich and precise manipulation.
-
InDex: Empowering VLA Models with Intent-Conditioned Arm-Hand Coordination for Dexterous Manipulation
InDex adapts parallel-gripper VLA policies to dexterous hands by conditioning a diffusion hand policy on a normalized scalar grasp intent, raising average simulated task success from 50.3% to 85.8%.
-
Efficient Skill Grounding via Code Refactoring with Small Language Models
RECENT decouples skill semantics from embodiment-specific bindings via code refactoring to let small language models achieve skill grounding performance matching large language model baselines.
-
MonoDuo: Using One Robot Arm to Learn Bimanual Policies
MonoDuo generates synthetic bimanual demonstrations from single-arm teleoperation plus human collaboration to train policies achieving up to 70% zero-shot success on five manipulation tasks, with 65-70% gains from 25-shot finetuning.
-
A Mechanistic Analysis of Sim-and-Real Co-Training in Generative Robot Policies
Sim-and-real co-training for robot policies is driven primarily by balanced cross-domain representation alignment and secondarily by domain-dependent action reweighting.
-
mimic-video: Video-Action Models for Generalizable Robot Control Beyond VLAs
mimic-video combines internet video pretraining with a flow-matching decoder to achieve state-of-the-art robotic manipulation performance with 10x better sample efficiency than vision-language-action models.
-
DreamPolicy: A Unified World-model Policy for Scalable Humanoid Locomotion
DreamPolicy integrates an autoregressive diffusion world model with policy learning to produce a single scalable policy that generalizes to unseen composite terrains for humanoid locomotion.
-
KITE: Decoupling Kinematics and Interaction for Zero-Shot Cross-Embodiment Manipulation
KITE decouples task reasoning from embodiment-specific control via learned latent interaction intents to enable zero-shot transfer across structurally different robots.
-
GR-3 Technical Report
GR-3 is a VLA model that generalizes to novel objects, environments, and abstract instructions, outperforms the π0 baseline, and integrates with the new ByteMini bi-manual mobile robot.