【李宏毅 MLDS18 国语】机器学习 深度学习 GAN 深度强化学习
http://speech.ee.ntu.edu.tw/~tlkagk/courses_MLDS18.html
作者：李宏毅（台湾大学）
时间：2018.03~2018.06
原链接：http://speech.ee.ntu.edu.tw/~tlkagk/courses_MLDS18.html
原网站有配套的PPT以及HomeWork.
视频选集
（28/35）
自动连播
Deep Learning Theory 1-1_ Can shallow network fit any function_
36:49
Deep Learning Theory 1-2_ Potential of Deep
56:09
Deep Learning Theory 1-3_ Is Deep better than Shallow_
26:18
Deep Learning Theory 2-1_ When Gradient is Zero
01:17:29
Deep Learning Theory 2-2_ Deep Linear Network
31:29
Deep Learning Theory 2-3_ Does Deep Network have Local Minima_
17:40
Deep Learning Theory 2-4_ Geometry of Loss Surfaces (Conjecture)
17:06
Deep Learning Theory 2-5_ Geometry of Loss Surfaces (Empirical)
38:46
Deep Learning Theory 3-1_ Generalization Capability of Deep Learning
25:38
Deep Learning Theory 3-2_ Indicator of Generalization
42:56
Computational Graph & Backpropagation
01:00:54
Sequence-to-sequence Learning
01:52:50
Pointer Network
13:34
Recursive Network
24:58
Attention-based Model
35:40
ForDeep
01:09:42
Tuning Hyperparameters
23:43
GAN Lecture 1 (2018)_ Introduction
01:33:15
GAN Lecture 2 (2018)_ Conditional Generation
26:19
GAN Lecture 3 (2018)_ Unsupervised Conditional Generation
38:59
GAN Lecture 4 (2018)_ Basic Theory
01:20:19
GAN Lecture 5 (2018)_ General Framework
25:04
GAN Lecture 6 (2018)_ WGAN, EBGAN
50:07
GAN Lecture 7 (2018)_ Info GAN, VAE-GAN, BiGAN
46:03
GAN Lecture 8 (2018)_ Photo Editing
22:46
GAN Lecture 9 (2018)_ Sequence Generation
01:27:24
GAN Lecture 10 (2018)_ Evaluation u0026 Concluding Remarks
30:09
DRL Lecture 1_ Policy Gradient (Review)
45:49
DRL Lecture 2_ Proximal Policy Optimization (PPO)
41:34
DRL Lecture 3_ Q-learning (Basic Idea)
49:44
DRL Lecture 4_ Q-learning (Advanced Tips)
38:31
DRL Lecture 5_ Q-learning (Continuous Action)
14:58
DRL Lecture 6_ Actor-Critic
34:16
DRL Lecture 7_ Sparse Reward
30:16
DRL Lecture 8_ Imitation Learning
34:02
顶部