CS - Lecture 13 - Self-Supervised Learning

Generative vs Self-Supervised Learning

• Both aim to learn from data without manual label annotation.

• Self-supervised Learning solves "pretext" task that produce good features for downstream tasks.

  1. Learn with supervised learning objectives (classification, regression)

  2. Labels of these pretex tasks are generated automatically.

Pretext Task

• Solving the pretext tasks allow the model to learn good features.

• We can automatically generate labels for the pretext tasks.

Pretext Task

• Solving the pretext tasks allow the model to learn good features.

• We can automatically generate labels for the pretext tasks.