机器学习基石

介绍各领域中的机器学习使用者都应该知道的基础算法、理论及实用工具

• 课时1：Course Introduction (10分58秒)

• 课时2：What is Machine Learning (18分28秒)

• 课时3：Applications of Machine Learning (18分56秒)

• 课时4：Components of Machine Learning (11分45秒)

• 课时5：Machine Learning and Other Fields (10分21秒)

• 课时6：Perceptron Hypothesis Set (15分42秒)

• 课时7：Perceptron Learning Algorithm (PLA) (19分46秒)

• 课时8：Guarantee of PLA (12分37秒)

• 课时9：Non-Separable Data (12分55秒)

• 课时10：Learning with Different Output Space (17分26秒)

• 课时11：Learning with Different Data Label (18分12秒)

• 课时12：Learning with Different Protocol (11分9秒)

• 课时13：Learning with Different Input Space (14分13秒)

• 课时14：Learning is Impossible (13分32秒)

• 课时15：Probability to the Rescue (11分33秒)

• 课时16：Connection to Learning (16分46秒)

• 课时17：Connection to Real Learning (18分6秒)

• 课时18：Recap and Preview (13分44秒)

• 课时19：Effective Number of Lines (15分26秒)

• 课时20：Effective Number of Hypotheses (16分17秒)

• 课时21：Break Point (7分44秒)

• 课时22：Restriction of Break Point (14分18秒)

• 课时23：Bounding Function- Basic Cases (6分56秒)

• 课时24：Bounding Function- Inductive Cases (14分47秒)

• 课时25：A Pictorial Proof (16分1秒)

• 课时26：Definition of VC Dimension (13分10秒)

• 课时27：VC Dimension of Perceptrons (13分27秒)

• 课时28：Physical Intuition of VC Dimension (6分11秒)

• 课时29：Interpreting VC Dimension (17分13秒)

• 课时30：Noise and Probabilistic Target (17分1秒)

• 课时31：Error Measure (15分10秒)

• 课时32：Algorithmic Error Measure (13分46秒)

• 课时33：Weighted Classification (16分54秒)

• 课时34：Linear Regression Problem (10分8秒)

• 课时35：Linear Regression Algorithm (20分2秒)

• 课时36：Generalization Issue (20分34秒)

• 课时37：Linear Regression for Binary Classification (11分23秒)

• 课时38：Logistic Regression Problem (14分33秒)

• 课时39：Logistic Regression Error (15分58秒)

• 课时40：Gradient of Logistic Regression Error (15分38秒)

• 课时41：Gradient Descent (19分18秒)

• 课时42：Linear Models for Binary Classification (21分35秒)

• 课时43：Stochastic Gradient Descent (11分39秒)

• 课时44：Multiclass via Logistic Regression (14分18秒)

• 课时45：Multiclass via Binary Classification (11分35秒)

• 课时46：Quadratic Hypothesis (23分47秒)

• 课时47：Nonlinear Transform (9分52秒)

• 课时48：Price of Nonlinear Transform (15分37秒)

• 课时49：Structured Hypothesis Sets (9分36秒)

• 课时50：What is Overfitting (10分45秒)

• 课时51：The Role of Noise and Data Size (13分36秒)

• 课时52：Deterministic Noise (14分7秒)

• 课时53：Dealing with Overfitting (10分49秒)

• 课时54：Regularized Hypothesis Set (19分16秒)

• 课时55：Weight Decay Regularization (24分8秒)

• 课时56：Regularization and VC Theory (8分15秒)

• 课时57：General Regularizers (13分28秒)

• 课时58：Model Selection Problem (16分0秒)

• 课时59：Validation (13分24秒)

• 课时60：Leave-One-Out Cross Validation (16分6秒)

• 课时61：V-Fold Cross Validation (10分41秒)

• 课时62：Occam-'s Razor (10分8秒)

• 课时63：Sampling Bias (11分50秒)

• 课时64：Data Snooping (12分28秒)

• 课时65：Power of Three (8分49秒)