This can be both a starting point for the development of artificial intelligence (including artificial neural network) algorithms, and a handbook containing introductory to advanced tools and solutions in this field. I have recently started documenting everything I have learned (and am learning) in the area. Soon I'll develope it much more.

Overall Review of Context

Explanations about the context of each directory:

First of all, my theoretic summaries are placed in the notes/. There are different text files each subjected to a specific topic with classification of the context in a table of contents at the beginning of each.

Elementary coding instructions based on different platforms are are placed in the samples/ directory.If you are not much familiar to the basic concepts, in samples/, there are a bunch of code examples related to different known ANN developement platforms. So the scripts are classified based on the different platforms. For instance, I suggest the samples/tensorflow/keras/ for a basic introduction to the artificial neural network design and implementation.

After all, the main helpful sample projects exist in projects/. Despite the samples/ directory, the scripts are classified based on the main topic of each. Withing the sample codes, baisc instruction subjected to the specific topic are given in comments.

The references for this context are a lot. I've tried to give a link to the reference in each code or note.

Learning Roadmap

The following is the best order for one to study the stuff. Actually this a table of contents ordered such that I personally have learned the stuff. I mean, this order is just a suggestion. Do whatever you like!

1. MLP: Theory - Samples

2. Custom Networks in Keras: Samples

3. Unsupervised Learning with SOM: Theory

4. DRL: Theory - Samples

Learning Roadmap: DRL

For DRL documents, go to directory: notes/DRL/. The order of documents (.md files) is as follows:

1. Fundamentals
2. DQN
3. PG

After reviewing the notes, for DRL scripts, go to directory: samples/DRL/.The order of codes (.py files) is as follows. This is how I dicumented the codes and the code comments are continued in this order:

1. PG Agent (tf2) (---> network - training) (You can also take a look at the same sample in keras)

2. DQN Agent (tf2) (---> trainging)

3. DQN Agent with CNN (keras) (Make sure to see the trainging code)

4. DQN Agent (keras) (---> trainging code)

5. DDQN Agent (keras) (---> trainging code)

6. Dueling DDQN Agent (tf2) (---> trainging code)