Skip to content
AI.info

AI.info

How machines learn

Twenty-six lessons on what actually changes with experience — from a real problem to a learning task you can defend.

  1. Learning in AI: What Changes With Experience?
  2. From a Real Problem to a Learning Task
  3. Examples, Observations, and the Unit of Experience
  4. Labels, Targets, and the Limits of Ground Truth
  5. Features, Representations, and What the Model Can See
  6. Parameters, Hyperparameters, and Model Capacity
  7. Building a Dataset That Represents the Job
  8. Training, Validation, and Test: Three Different Jobs
  9. Data Leakage: When the Answer Sneaks Into the Question
  10. The Training Loop: Predict, Measure, Update
  11. Loss Functions: Designing the Feedback Signal
  12. Gradients, Batches, Epochs, and Learning Rate
  13. Baselines and Sanity Checks
  14. Underfitting: When the Model Cannot Learn Enough
  15. Overfitting: When the Model Learns the Dataset Too Closely
  16. Generalization and Inductive Bias
  17. Regularization, Early Stopping, and Capacity Control
  18. Metrics, Thresholds, and the Cost of Different Errors
  19. Error Analysis: Learn From the Examples the Model Misses
  20. Data Quality, Class Imbalance, and Label Noise
  21. Learning Curves and When More Data Helps
  22. Distribution Shift: When the World Stops Matching the Dataset
  23. Feedback Loops and Selective Labels
  24. Experiments, Randomness, and Reproducibility
  25. A Diagnosis-First Strategy for Improving Models
  26. Capstone: Design an Honest Learning Experiment