Self-Supervised Learning
AI/ML Fundamentals
Learning useful representations by creating proxy tasks from raw data
What is Self-Supervised Learning?
Models learn from unlabeled data by predicting parts of input (e.g., masked language modeling) then fine-tune on downstream tasks.
Real-World Examples
- •BERT pretraining
- •Contrastive learning for images
Related Terms
Learn more about concepts related to Self-Supervised Learning