机器学习是一门让计算机在非精确编程下進行活动的科学。在过去十年，机器学习促成了无人驾驶车、高效语音识别、精确网络搜索及人类基因组认知的大力发展。机器学习如此无孔不入，你可能已经在不知情的情况下利用过无数次。许多研究者认为，这种手段是达到人类水平AI的最佳方式。这门课程中，你将学习到高效的机器学习技巧，及学会如何利用它为你服务。重点是，你不仅能学到理论基础，更能知晓如何快速有效应用这些技巧到新的问题上。最后，你会接触到硅谷创新中几个优秀的涉及机器学习与AI的应用实例。

此课程将广泛介绍机器学习、数据挖掘与统计模式识别的知识。

主题包括：

(i) 监督学习（参数/非参数算法、支持向量机、内核、神经网络）。

(iii) 机器学习的优秀案例（偏差/方差理论；机器学习和人工智能的创新过程）课程将拮取案例研究与应用，学习如何将学习算法应用到智能机器人（观感，控制）、文字理解（网页搜索，防垃圾邮件）、计算机视觉、医学信息学、音频、数据挖掘及其他领域上。

【课程内容】

1 – 1 – Welcome (7 min)

1 – 2 – What is Machine Learning- (7 min)

1 – 3 – Supervised Learning (12 min)

1 – 4 – Unsupervised Learning (14 min)

2 – 1 – Model Representation (8 min)

2 – 2 – Cost Function (8 min)

2 – 3 – Cost Function – Intuition I (11 min)

2 – 4 – Cost Function – Intuition II (9 min)

2 – 5 – Gradient Descent (11 min)

2 – 6 – Gradient Descent Intuition (12 min)

2 – 7 – Gradient Descent For Linear Regression (10 min)

2 – 8 – What-'s Next (6 min)

3 – 1 – Matrices and Vectors (9 min)

3 – 2 – Addition and Scalar Multiplication (7 min)

3 – 3 – Matrix Vector Multiplication (14 min)

3 – 4 – Matrix Matrix Multiplication (11 min)

3 – 5 – Matrix Multiplication Properties (9 min)

3 – 6 – Inverse and Transpose (11 min)

4 – 1 – Multiple Features (8 min)

4 – 2 – Gradient Descent for Multiple Variables (5 min)

4 – 3 – Gradient Descent in Practice I – Feature Scaling (9 min)

4 – 4 – Gradient Descent in Practice II – Learning Rate (9 min)

4 – 5 – Features and Polynomial Regression (8 min)

4 – 6 – Normal Equation (16 min)

4 – 7 – Normal Equation Noninvertibility (Optional) (6 min)

5 – 1 – Basic Operations (14 min)

5 – 2 – Moving Data Around (16 min)

5 – 3 – Computing on Data (13 min)

5 – 4 – Plotting Data (10 min)

5 – 5 – Control Statements- for, while, if statements (13 min)

5 – 6 – Vectorization (14 min)

5 – 7 – Working on and Submitting Programming Exercises (4 min)

6 – 1 – Classification (8 min)

6 – 2 – Hypothesis Representation (7 min)

6 – 3 – Decision Boundary (15 min)

6 – 4 – Cost Function (11 min)

6 – 5 – Simplified Cost Function and Gradient Descent (10 min)

6 – 6 – Advanced Optimization (14 min)

6 – 7 – Multiclass Classification- One-vs-all (6 min)

7 – 1 – The Problem of Overfitting (10 min)

7 – 2 – Cost Function (10 min)

7 – 3 – Regularized Linear Regression (11 min)

7 – 4 – Regularized Logistic Regression (9 min)

8 – 1 – Non-linear Hypotheses (10 min)

8 – 2 – Neurons and the Brain (8 min)

8 – 3 – Model Representation I (12 min)

8 – 4 – Model Representation II (12 min)

8 – 5 – Examples and Intuitions I (7 min)

8 – 6 – Examples and Intuitions II (10 min)

8 – 7 – Multiclass Classification (4 min)

9 – 1 – Cost Function (7 min)

9 – 2 – Backpropagation Algorithm (12 min)

9 – 3 – Backpropagation Intuition (13 min)

9 – 4 – Implementation Note- Unrolling Parameters (8 min)

9 – 5 – Gradient Checking (12 min)

9 – 6 – Random Initialization (7 min)

9 – 7 – Putting It Together (14 min)

9 – 8 – Autonomous Driving (7 min)

10 – 1 – Deciding What to Try Next (6 min)

10 – 2 – Evaluating a Hypothesis (8 min)

10 – 3 – Model Selection and Train-Validation-Test Sets (12 min)

10 – 4 – Diagnosing Bias vs. Variance (8 min)

10 – 5 – Regularization and Bias-Variance (11 min)

10 – 6 – Learning Curves (12 min)

10 – 7 – Deciding What to Do Next Revisited (7 min)

11 – 1 – Prioritizing What to Work On (10 min)

11 – 2 – Error Analysis (13 min)

11 – 3 – Error Metrics for Skewed Classes (12 min)

11 – 4 – Trading Off Precision and Recall (14 min)

11 – 5 – Data For Machine Learning (11 min)

12 – 1 – Optimization Objective (15 min)

12 – 2 – Large Margin Intuition (11 min)

12 – 3 – Mathematics Behind Large Margin Classification (Optional) (20 min)

12 – 4 – Kernels I (16 min)

12 – 5 – Kernels II (16 min)

12 – 6 – Using An SVM (21 min)

13 – 1 – Unsupervised Learning- Introduction (3 min)

13 – 2 – K-Means Algorithm (13 min)

13 – 3 – Optimization Objective (7 min)

13 – 4 – Random Initialization (8 min)

13 – 5 – Choosing the Number of Clusters (8 min)

14 – 1 – Motivation I- Data Compression (10 min)

14 – 2 – Motivation II- Visualization (6 min)

14 – 3 – Principal Component Analysis Problem Formulation (9 min)

14 – 4 – Principal Component Analysis Algorithm (15 min)

14 – 5 – Choosing the Number of Principal Components (11 min)

14 – 6 – Reconstruction from Compressed Representation (4 min)

14 – 7 – Advice for Applying PCA (13 min)

15 – 1 – Problem Motivation (8 min)

15 – 2 – Gaussian Distribution (10 min)

15 – 3 – Algorithm (12 min)

15 – 4 – Developing and Evaluating an Anomaly Detection System (13 min)

15 – 5 – Anomaly Detection vs. Supervised Learning (8 min)

15 – 6 – Choosing What Features to Use (12 min)<br helvetica="" neue",="" helvetica,="" arial,="" "pingfang="" sc",="" "hiragino="" sans="" gb",="" stheiti,="" "microsoft="" yahei",="" jhenghei",="" simsun,="" sans-serif;="" font-size:="" 14px;"="">