About Explicit Variance Minimization: Training Neural Networks for Medical Imaging With Limited Data Annotations

1 code implementation • 28 May 2021 • Dmitrii Shubin, Danny Eytan, Sebastian D. Goodfellow

Self-supervised learning methods for computer vision have demonstrated the effectiveness of pre-training feature representations, resulting in well-generalizing Deep Neural Networks, even if the annotated data are limited.

Representation Learning Self-Supervised Learning +2