Neural Networks for Machine Learning From Scratch (Udemy.com)

Awesome course!!!

It is an effective learning course for those who want to start NN, because the necessary basic information is supported by a nice simple language and appropriate presentation.

Audio quality is too bad

Pros: hands-on programming from scratch

Cons: thick accent and audio

The coding is great, but the explanations

Excellent job!

As a person, new to neural networks, I found the information given were satisfying. You can start easily enough with the given course material. It was more advanced than Tensorflow - Deep Learning course but it was still comprehensible. Educator's support after the course was something which made the training satisfactory.

successful, fluent, understandable and very good at sampling

The course contains simple and clear content to teach people at all levels. Very successful!

This course is direct to the point and shows us what is behind the curtain of neural networks apis. Even you will just use popular framework and their apis instead of writing your own implementation. It is good to understand how they are implemented. I would highly recommend this course if you are curious about what is going on behind the scene.

The instructor is clearly knowledgable about the subject, and gets straight into coding. The course is centered around building a basic program in the first lectures, and then expands on this program in the following ones. If you like to get your hands straight into some code and see some results, this is a really nice approach. Just beware that the coding pace is sometimes at a bit high, so expect to pause the video from time to time to catch up.

The main thing I'm missing is some details bridging the mathematics (the instructor has posted these on an external source - probably preferable to videos for reading through formulas) and the code. I can see that the code works, and I have an overall idea of what's going on, but I don't quite feel confident that I would be able to explain every bit of the code.

In this course, you can really learn how to construct nodes, make connections between synaptic weights then you can learn feedforward and backpropagation algorithms in practical ways. These topics are explained well. You can also have ideas about the math that is behind of the algorithms, but sometimes you may have hard time to understand what's going on, because the instructor wrote down the codes in a lecture after that lecture when you look at his codes in the following lecture videos, there are some changes that he never explained why did he change the codes. On the other hand, he needs to explain every line that he codes. He doesn't explain every line, but you get the main idea. This is a suggestion to him he should ask himself that "why do we write this line? What is aim of this line" etc. If you ask many questions then the things become easy for you.