Machine heart editor
Participation: Siyuan, Zenan
Ntu professor Li Hongyi’s machine learning course is often cited as the top choice for Chinese open courses. Professor Li’s teaching style is humorous and easy to understand. His courses not only cover the basic knowledge of machine learning and deep learning, but also introduce various latest technologies in THE ML field. Recently, the 2019 course materials and videos are finally online!
-
The course information link: http://speech.ee.ntu.edu.tw/~tlkagk/courses_ML19.html
-
Course video (Bilibili) : https://www.bilibili.com/video/av46561029/
-
YouTube link: https://www.youtube.com/playlist?list=PLJV_el3uVTsOK_ZK5L0Iv_EQoL1JefRL4
Hung-yi Lee is currently an Assistant Professor in the Department of Electrical Engineering and The School of Electrical And Information Engineering at National Taiwan University. He received his PhD from National Taiwan University in 2012 and was a visiting scholar at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT) in 2013. His main research interests are machine learning (deep learning) and speech recognition.
His video for the 2016 machine Learning course was a popular learning material.
It’s not easy to have a professor who understands quadratic elements.
The course catalog of machine Learning 2019, in parentheses are the new contents:
-
Regression, gradient descent
-
Classification, logistic regression, the cause of misclassification
-
Deep learning, back propagation (exception detection)
-
Convolutional Neural Networks, Keras (Counter sample Attack)
-
Training deep Learning models (to Explain AI)
-
Cyclic Neural Network (Order LSTM)
-
Ensemble
-
Semi-supervised learning, transfer learning (lifelong learning)
-
(Meta-learning)
-
Seq2seq (Transformer)
-
(Few/Zero Shot Learning)
-
Unsupervised learning (BERT)
-
Reinforcement learning (more detailed)
-
(Network compression)
-
Generative Adversarial Network (GLOW)
-
(Unsupervised domain adaptation)
-
Why use deep Learning (Deep learning theory)
Those of you who have watched Li hongyi’s fall 2017 machine learning course know that he is very careful about the basics. For example, for the recurrent neural network, he will take us through manual operation, so as to figure out the input of each time step, the memory stored and the specific operation process, etc. In 19 years of new lessons, Li Hongyi focused on opening videos and assignments related to new lessons.
Most of the new courses are recently popular research frontiers, such as Full attention Network Transformer in Seq2Seq and the recently popular new paradigm flow Model (Glow). These can complement the previous 17 fall courses, bringing the video as a whole closer to the current edge.
At present, Li Hongyi has released videos of anomaly detection and anti-attack, which are newly added content. These additions are best seen in conjunction with the main course so they can be better understood. For example, how can exception detection be seen together with deep learning foundations, countermeasures against attacks can be seen together with convolutional neural networks and so on.
The above is a screenshot of the YouTube video. Students who have not logged on to the Internet can also directly see the resources uploaded by Aikeke teacher to station B. Of course, in addition to a series of newly updated courses on YouTube, Professor Li Hongyi has published more course topics, such as linear algebra, deep reinforcement learning, generative adversarial networks, deep learning theory, machine learning (Fall of ’17), etc.
Machine Learning in The fall of 17 gives a general overview of machine learning and deep learning, which is also a must-see part of Li Hongyi’s course. After learning the machine learning course, we basically have a certain understanding of various topics, so we can further look at his views on advanced topics of deep learning, generative adversarial networks and so on. The following is a list of topics and videos opened by Li Hongyi:
Finally, watching videos and doing homework requires persistence. I hope everyone can turn these resources into their own knowledge.
This article is edited by machine heart. Please contact this official account for reprinting.
✄ — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
Join Heart of the Machine (full-time reporter/intern) : [email protected]
Contribute or seek coverage: [email protected]
Advertising & Business partnerships: [email protected]