Machine learningMachine learning

Samonadzorirano aktivno učenje

Samonadzorirano aktivno učenje (SSL-AL) je paradigm strokovnog učenja učinkovito u pogledu oznaka koji prethodno trenira model na neoznačenim podacima koristeći samonadzorirane ciljeve, a zatim strateški upućuje upite ljudskom orakulu za najinformativnije oznake koristeći funkciju stjecanja aktivnog učenja. Rezultat je snažna prediktivna izvedba s djelićem troškova anotacije potrebnih za potpuno nadzirane pristupe.

Otvorite u MethodMindUskoroVideoUskoroDownload slides

Pročitajte cijelu metodu

Samo za članove

Prijavite se besplatnim računom kako biste pročitali ovaj odjeljak.

Prijavite se

Method map

The neighbourhood of related methods — select a node to explore.

Izvori

  1. Bengar, J. Z., van de Weijer, J., Twardowski, B., & Raducanu, B. (2021). Reducing Label Effort: Self-Supervised Meets Active Learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), pp. 1631–1639. link
  2. Zhan, X., Wang, Q., Huang, K.-H., Xiong, H., Dou, D., & Chan, A. B. (2022). A comparative survey of deep active learning. arXiv preprint arXiv:2203.13450. link

Kako citirati ovu stranicu

ScholarGate. (2026, June 3). Self-supervised Active Learning (SSL-AL hybrid label-efficient framework). ScholarGate. https://scholargate.app/hr/machine-learning/self-supervised-active-learning

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

Compare side by side
ScholarGateSelf-supervised Active Learning (Self-supervised Active Learning (SSL-AL hybrid label-efficient framework)). Preuzeto 2026-06-15 s https://scholargate.app/hr/machine-learning/self-supervised-active-learning · Skup podataka: https://doi.org/10.5281/zenodo.20539026