Masked Language Modeling

Masked Language Modeling is a self-supervised learning method in natural language processing where the model is trained to predict randomly masked tokens in a sequence based on their contextual representations, distinguishing it from other language modeling approaches by its bidirectional context utilization and explicit reconstruction objective.

411

Publications

34.9K

Citations

1.6K

Authors

345

Institutions

Publications per year

2017–2026

411

Authors

1.6K

  1. LZ
    Luke Zettlemoyer

    University of Washington, United States

    16 publications · 1.4K citations

    15

    h-index

  2. ZG
    Zhe Gan

    Microsoft Research (United Kingdom), United Kingdom

    10 publications · 1.1K citations

    10

    h-index

  3. MG
    Marjan Ghazvininejad

    Meta (Israel), Israel

    9 publications · 421 citations

    9

    h-index

  4. ND
    Nan Duan

    Microsoft Research Asia (China), China

    9 publications · 1.9K citations

    9

    h-index

  5. XQ
    Xipeng Qiu

    Fudan University, China

    7 publications · 737 citations

    7

    h-index

1–5 of 1.6K

Venues

  1. arXiv (Cornell University)

    Repository · 122 publications · 7.8K citations

    40

    h-index

  2. 12

    h-index

  3. 12

    h-index

  4. Proceedings of the AAAI Conference on Artificial Intelligence

    Conference · 11 publications · 1.1K citations

    11

    h-index

  5. bioRxiv (Cold Spring Harbor Laboratory)

    Repository · 7 publications · 558 citations

    7

    h-index

1–5