en-es
en-fr
en-sl
en
0.25
0.5
0.75
1.25
1.5
1.75
2
Machine Learning
Published on Aug 23, 201652070 Views
We provide a general introduction to machine learning, aimed to put all participants on the same page in terms of definitions and basic background. After a brief overview of different machine learning
Related categories
Chapter list
Introduction to Machine Learning00:00
Outline00:13
Types of machine learning problems00:53
Supervised learning01:08
Example: Face detection and recognition02:06
Reinforcement learning02:23
Example: TD-Gammon (Tesauro, 1990-1995)03:06
Unsupervised learning03:32
Example: Oncology (Alizadeh et al.)04:08
Example: A data set04:31
Data (continued)05:07
Terminology05:47
More formally06:23
Supervised learning problem06:48
Steps to solving a supervised learning problem07:07
Example: What hypothesis class should we pick?07:22
Linear hypothesis07:30
Error minimization!07:56
Least mean squares (LMS)08:22
Steps to solving a supervised learning problem08:56
Notation reminder09:23
A bit of algebra09:51
The solution10:09
Example: Data and best linear hypothesis11:02
Linear regression summary11:03
Linear function approximation in general11:05
Linear models in general12:18
Remarks12:28
Order-2 fit13:43
Order-4 fit14:18
Order-5 fit14:20
Order-6 fit14:21
Order-7 fit14:22
Order-8 fit14:24
Order-3 fit15:33
Order-9 fit16:17
Overfitting16:26
Overfitting and underfitting17:01
Overfitting more formally18:16
Typical overfitting plot19:24
Cross-validation21:59
The anatomy of the error of an estimator22:34
Bias-variance analysis24:02
Recall: Statistics 10124:54
The variance lemma25:13
Bias-variance decomposition25:22
Bias-variance decomposition (2)27:26
Error decomposition29:34
Bias-variance trade-off30:52
More on overfitting31:51
Coming back to mean-squared error function...32:34
A probabilistic assumption33:38
Bayes theorem in learning34:36
Choosing hypotheses35:32
Maximum likelihood estimation36:20
The log trick37:26
Maximum likelihood for regression37:51
Applying the log trick38:06
Maximum likelihood hypothesis for least-squares estimators39:57
A graphical representation for the data generation process44:39
Regularization44:54
Regularization for linear models46:14
What L2 regularization does47:32
L2 Regularization for linear models revisited48:00
Visualizing regularization (2 parameters)48:36
Pros and cons of L2 regularization49:46
Visualizing L1 regularization50:36
L1 Regularization for linear models51:42
Solving L1 regularization52:02
Pros and cons of L1 regularization52:30
Example of L1 vs L2 effect52:55
Bayesian view of regularization - 153:29
Bayesian view of regularization - 255:35
What does the Bayesian view give us? - 159:11
What does the Bayesian view give us? - 201:01:36
What does the Bayesian view give us? - 301:02:09
Logistic regression01:02:54
The cross-entropy error function01:04:40
Cross-entropy error surface for logistic function01:05:45
Gradient descent01:06:16
Example gradient descent traces01:07:04
Gradient descent algorithm01:07:14
Maximization procedure: Gradient ascent01:08:14
Another algorithm for optimization01:08:42
Application to machine learning01:10:10
Second-order methods: Multivariate setting01:11:06
Which method is better?01:11:46
Newton-Raphson for logistic regression01:12:21
Regularization for logistic regression01:12:49
Probabilistic view of logistic regression01:14:01
Recap01:16:00