The classifier judged this item not relevant to AI security, so it is left out of lists, feeds, the API and dataset releases. If that is wrong, report a correction on this page.
InfoResearchPeer-reviewed
Human Behavior Anonymization for Secure Teleoperation
- Published
- Record updated
Summary
Researchers propose a disentangled representation-learning framework based on a Variational Autoencoder (VAE) to suppress identity-revealing cues in teleoperation motion streams while keeping task-relevant motion patterns. Evaluated offline on reconstructed trajectories from a tele-robotic prototype with multiple users performing manipulation tasks, the approach substantially reduces re-identification risk and offers a favorable privacy-utility trade-off in task utility.