REVIEW 7 cited by
AirExo-2: Scaling up Generalizable Robotic Imitation Learning with Low-Cost Exoskeletons
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
AirExo-2: Scaling up Generalizable Robotic Imitation Learning with Low-Cost Exoskeletons
read the original abstract
Scaling up robotic imitation learning for real-world applications requires efficient and scalable demonstration collection methods. While teleoperation is effective, it depends on costly and inflexible robot platforms. In-the-wild demonstrations offer a promising alternative, but existing collection devices have key limitations: handheld setups offer limited observational coverage, and whole-body systems often require fine-tuning with robot data due to domain gaps. To address these challenges, we present AirExo-2, a low-cost exoskeleton system for large-scale in-the-wild data collection, along with several adaptors that transform collected data into pseudo-robot demonstrations suitable for policy learning. We further introduce RISE-2, a generalizable imitation learning policy that fuses 3D spatial and 2D semantic perception for robust manipulations. Experiments show that RISE-2 outperforms prior state-of-the-art methods on both in-domain and generalization evaluations. Trained solely on adapted in-the-wild data produced by AirExo-2, the RISE-2 policy achieves comparable performance to the policy trained with teleoperated data, highlighting the effectiveness and potential of AirExo-2 for scalable and generalizable imitation learning.
Forward citations
Cited by 7 Pith papers
-
DexDirect: Direct Kinesthetic Arm Guidance for Efficient Dexterous Demonstration Collection
A hybrid kinesthetic-arm-plus-webcam-hand teleoperation interface achieved 17x/3x higher demonstration throughput than vision baselines and trained a 90%-success pick-and-place policy in a ten-person study.
-
HiFi-UMI: Learning Deployable Manipulation Policies from High-Fidelity UMI Data Alone
Robot-free HiFi-UMI demonstrations can replace teleoperated real-robot data in post-training: three policy backbones matched in-domain teleoperation within 3.1 percentage points, including 85% success on a precision i...
-
ModPack: An Extensible Teleoperation Interface for Bimanual Mobile Manipulation
A modular backpack-based teleoperation interface enables bimanual mobile manipulation with haptic feedback and active perception across multiple robot platforms.
-
Unify Robot Actions in Camera Frame
CalibAll estimates camera extrinsics on existing datasets to convert robot actions into a unified camera-frame representation, enabling stronger cross-embodiment pretraining.
-
KAI: A Kinematic-Aware Interface for Data-Efficient Articulated Object Manipulation
KAI, a keypoint-and-displacement intermediate with geometric joint priors, matches or beats articulated-manipulation baselines at half the demo data and supports human-video co-training.
-
R3D: Revisiting 3D Policy Learning
A transformer 3D encoder plus diffusion decoder architecture, with 3D-specific augmentations, outperforms prior 3D policy methods on manipulation benchmarks by improving training stability.
-
Human Motion Data Alone Does Not Guarantee Plausible Gait Biomechanics
Motion-only imitation learning reproduces walking kinematics but produces inaccurate ground reaction forces and joint moments; adding GRF and center-of-pressure rewards brings simulated kinetics closer to inverse dynamics.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.