機器學習基石下 (Machine Learning Foundations)---Algorithmic Foundations

Module 1: Linear Regression

• Explore the linear regression problem and algorithm
• Understand generalization issues and its application for binary classification
• Get insights into the course outline, format, grading criteria, and extended reading

Module 2: Logistic Regression

• Learn about the logistic regression problem, error, gradient, and gradient descent

Module 3: Linear Models for Classification

• Understand linear models for binary classification, stochastic gradient descent, and multiclass classification via logistic regression and binary classification

Module 4: Nonlinear Transformation

• Explore quadratic hypothesis, nonlinear transformation, price of nonlinear transform, and structured hypothesis sets
• Complete assignment three

Module 5: Hazard of Overfitting

• Understand overfitting, the role of noise and data size, deterministic noise, and techniques for dealing with overfitting

Module 6: Regularization

• Get insights into regularized hypothesis set, weight decay regularization, regularization and VC theory, and general regularizers

Module 7: Validation

• Explore the model selection problem, validation, leave-one-out cross-validation, and V-fold cross-validation

Module 8: Three Learning Principles

• Learn about Occam's Razor, sampling bias, data snooping, and the power of three
• Complete assignment four