Backbone Freezing

Backbone Freezing is a continual learning strategy in which the parameters of the foundational layers of a neural network are immobilized during subsequent task learning to preserve previously acquired knowledge while enabling adaptation in shallower or task-specific layers.

4

Publications

109

Citations

18

Authors

9

Institutions

Publications per year

2017–2026

4

Variants

Parameter FreezingWeight Freezing

Authors

18

  1. SH
    Si Hyun Han

    Sungkyunkwan University, South Korea

    1 publications · 36 citations

    1

    h-index

  2. WP
    Wan Pyo Hong

    Sungkyunkwan University, South Korea

    1 publications · 36 citations

    1

    h-index

  3. JG
    Ji Gwang Yu

    Sungkyunkwan University, South Korea

    1 publications · 36 citations

    1

    h-index

  4. JY
    Jun Yeob Lee

    Sungkyunkwan University, South Korea

    1 publications · 36 citations

    1

    h-index

  5. HL
    Ha Lim Lee

    Sungkyunkwan University, South Korea

    1 publications · 36 citations

    1

    h-index

1–5 of 18

Venues

  1. 2

    h-index

  2. Journal of Materials Chemistry C

    1 publications · 36 citations

    1

    h-index

  3. 1

    h-index

  4. arXiv (Cornell University)

    Repository · 1 publications · 13 citations

1–4