# Machine Learning, Probability and Graphical Models

Lecture slides:

- Probabilistic Graphical Models
- Building Intelligent Computers
- Uncertainty and Artificial Intelligence (UAI)
- Applications of Probabilistic Learning
- Canonical Tasks
- Representation
- Using random variables to represent the world
- Structure of Learning Machines
- Loss Functions for Tuning Parameters
- Training vs. Testing
- Sampling Assumption
- Generalization and Overfitting
- Capacity: Complexity of Hypothesis Space
- Inductive Bias
- Probabilistic Approach
- Joint Probabilities
- Conditional Independence
- Probabilistic Graphical Models
- Directed Graphical Models
- Example DAG
- Conditional Independence and Missing Edges in DAGs
- Explaining Away
- What's Inside the Nodes/Cliques?
- Probability Tables & CPTs
- Exponential Family
- Nodes with Parents
- Review: Goal of Graphical Models
- Learning Graphical Models from Data
- Basic Statistical Problems

*Author: Sam Roweis, Department Of Computer Science, University Of Toronto*