Aayush Talreja

Robot learning · Legged robots

  • StudyingComputer science and machine learning, University of Maryland
  • FocusWorld models from cross-embodiment video; autonomous robot data collection
  • AuRA: From Atoms to Bits through Autonomous Robot Interaction

    X. Yu*, Y. Jung*, H. Chen, Z. Cai, A. Talreja, S. Lee, H. Kim, F. Huang

    Robot-collected interaction data for physics-informed digital twins of deformable objects; 88.5% less human labor per asset.

    arXiv 2026PaperProjectCode

  • TraceGen: World Modeling in 3D Trace Space Enables Learning from Cross-Embodiment Videos

    S. Lee, Y. Jung, I. Chun, Y.-C. Lee, Z. Cai, H. Huang, A. Talreja, T. D. Dao, Y. Liang, J.-B. Huang, F. Huang

    A world model over 3D point traces instead of pixels, trained on human and cross-embodiment video.

    CVPR 2026PaperProjectCode

All writing →

EmailXLinkedIn