Skill Abstraction

Skill Abstraction is a methodological construct in self-supervised robot learning that identifies and encapsulates low-level sensorimotor patterns into reusable, high-level behavioral units to enable efficient generalization and transfer across tasks.

45

Publications

2.3K

Citations

138

Authors

50

Institutions

Publications per year

2017–2026

22

Authors

138

  1. AG
    Andrew G. Barto

    University of Massachusetts Amherst, United States

    5 publications · 591 citations

    5

    h-index

  2. GK
    George Konidaris

    University of Massachusetts Amherst, United States

    3 publications · 184 citations

    3

    h-index

  3. SG
    Shixiang Gu

    Google (United States), United States

    2 publications · 107 citations

    2

    h-index

  4. YJ
    Yiding Jiang

    Google (United States), United States

    2 publications · 107 citations

    2

    h-index

  5. GZ
    Grace Zhang

    Curtin University, Australia

    2 publications · 88 citations

    2

    h-index

1–5 of 138

Venues

  1. arXiv (Cornell University)

    Repository · 11 publications · 841 citations

    11

    h-index

  2. IEEE Transactions on Autonomous Mental Development

    TAMD · 2 publications · 116 citations

    2

    h-index

  3. IEEE Robotics and Automation Letters

    2 publications · 86 citations

    2

    h-index

  4. ScholarWorks@UMassAmherst (University of Massachusetts Amherst)

    Repository · 1 publications · 78 citations

    1

    h-index

  5. International Journal of Humanoid Robotics

    1 publications · 89 citations

    1

    h-index

1–5