Fundamentals of AI and ML
Learning Paradigms
Learning paradigms describe how a model receives training signals. The exam often tests whether labels, rewards, unlabeled data, or self-created objectives are used.
Concepts
- Supervised learning trains from labeled examples with known outputs.
- Unsupervised learning finds structure in unlabeled data.
- Reinforcement learning learns actions from rewards and penalties in an environment.
- Self-supervised learning creates training signals from the data itself and is common in foundation model pretraining.
Exam tips
- Supervised learning uses labeled examples with known answers.
- Unsupervised learning finds structure without labels.
- Reinforcement learning learns from rewards and penalties.