【深度学习】神经网络机器学习课程（Geoffrey Hinton 2012）（英文字幕）
https://www.youtube.com/playlist?list=PLoRl3Ht4JOcdU872GhiYWf6jwrk_SNhz9
Geoffrey Hinton经典神经网络课程
视频选集
（1/78）
自动连播
1.1 Why do we need machine learning
13:15
1.2 What are neural networks
08:30
1.3 Some simple models of neurons
08:24
1.4 A simple example of learning
05:39
1.5 Three types of learning
07:38
2.1 Types of neural network architectures
07:29
2.2 Perceptrons - first-generation neural networks
08:17
2.3 A geometrical view of perceptrons
06:25
2.4 Why the learning works
05:10
2.5 What perceptrons can't do
14:35
3.1 Learning the weights of a linear neuron
11:55
3.2 The error surface for a linear neuron
05:03
3.3 Learning weights of logistic output neuron
03:57
3.4 The backpropagation algorithm
11:51
3.5 Using the derivatives from backpropagation
09:50
4.1 Learning to predict the next word
12:34
4.2 A brief diversion into cognitive science
04:27
4.3 The softmax output function
07:21
4.4 Neuro-probabilistic language models
07:52
4.5 Dealing with many possible outputs
12:17
5.1 Why object recognition is difficult
04:40
5.2 Achieving viewpoint invariance
05:58
5.3 Convolutional nets for digit recognition
16:02
5.4 Convolutional nets for object recognition
17:44
6.1 Overview of mini batch gradient descent
08:23
6.2 A bag of tricks for mini batch gradient descent
13:16
6.3 The momentum method
08:43
6.4 Adaptive learning rates for each connection
05:44
6.5 Rmsprop - normalize the gradient
11:38
7.1 Modeling sequences - a brief overview
17:24
7.2 Training RNNs with back propagation
06:23
7.3 A toy example of training an RNN
06:15
7.4 Why it is difficult to train an RNN
07:44
7.5 Long term Short term memory
09:16
8.1 A brief overview of Hessian-free optimization
14:25
8.2 Modeling character strings
14:36
8.3 Predicting the next character using HF
12:25
8.4 Echo State Networks
09:38
9.1 Overview of ways to improve generalization
11:45
9.2 Limiting the size of the weights
06:23
9.3 Using noise as a regularizer
07:32
9.4 Introduction to the full Bayesian approach
10:50
9.5 The Bayesian interpretation of weight decay
10:53
9.6 MacKay's quick and dirty method
03:32
10.1 Why it helps to combine models
13:11
10.2 Mixtures of Experts
13:16
10.3 The idea of full Bayesian learning
07:28
10.4 Making full Bayesian learning practical
06:44
10.5 Dropout
08:35
11.1 Hopfield Nets
13:02
11.2 Dealing with spurious minima
11:03
11.3 Hopfield nets with hidden units
09:40
11.4 Using stochastic units to improve search
10:25
11.5 How a Boltzmann machine models data
11:45
12.1 Boltzmann machine learning
12:16
12.2 More efficient ways to get the statistics
14:49
12.3 Restricted Boltzmann Machines
10:54
12.4 An example of RBM learning
07:15
12.5 RBMs for collaborative filtering
08:17
13.1 The ups and downs of backpropagation
09:54
13.2 Belief Nets
12:36
13.3 Learning sigmoid belief nets
11:26
13.4 The wake sleep algorithm
13:15
14.1 Learning layers of features by stacking RBMs
17:35
14.2 Discriminative learning for DBNs
09:41
14.3 Discriminative fine tuning
08:39
14.4 Modeling real valued data with an RBM
09:56
14.5 RBMs are infinite sigmoid belief nets
17:12
15.1 From PCA to autoencoders
07:58
15.2 Deep autoencoders
04:11
15.3 Deep autoencoders for document retrieval
08:19
15.4 Semantic Hashing
08:50
15.5 Learning binary codes for image retrieval
09:38
15.6 Shallow autoencoders for pre-training
07:02
16.1 Learning a joint model of images and captions
09:05
16.2 Hierarchical Coordinate Frames
09:41
16.3 Bayesian optimization of hyper-parameters
13:30
16.4 The fog of progress
02:25
【B站AI论坛】欢迎写代码、做产品的AI开发者发帖加入
共建AI开发者交流地>>
顶部