Deep Reinforcement Learning

 The program is going great so far! The first project was very fun and educational. If you put serious effort into the projects and mini-projects, the reward is enormous. I am excited to plow forward! 

 Initially i was thinking am i able to get this first project complete -AI Navigator using Deep Q learning since mostly algorithms and some advanced concepts  but i made it in just 1 month itself because the way the concept explained with supportive code base and the course structure is awesome. Basically if we understand the concept right we can make it in any Language since only syntax is differ. Udacity Team made this Nano Degree in so unique way by making you understand the concept through detailed videos with more examples, analogy, quiz and provide code workspace through browser to make you actual hands-on with coding ,mentor support & knowledge base . My first project I completed using Udacity provided  workspace  even with additional GPU times (50 hrs),Thanks to the entire Udacity Team for your high quality , unique way of teaching and Keep en-lighting the aspirants! 

It is definitely the best online course on Reinforcement learning. The depth of the material nicely complements the ingenuity of the projects. 

I think the curriculum was excellent. The main instructor Alexis Cook is a brilliant teacher and describes the content extrmely well. However, same cannot be said of other instructors, especially those from the industry who taught DRL applications in finance were terrible. 

The projects were tougher than my previous Deep Learning Nanodegree and I learned a lot doing them. I got a pretty comprehensive view of DRL techniques through the three proijects. However, I feel there should have been one more project or at least an exercise on applying DRL on autonomous vehicles and Robotics to get a taste of that field.

Another suggestion is to make substantial changes in the projects in each cohort. As of now, many students simply search for project solutions posted by past students on Github and get a pretty good idea of the solution, and also use their code in bits and pieces at least (if not fully) without undergoing the rigorous experience the projects are aimed at providing.

Finally, I want to make a very important suggestion to Udacity guys. One key piece missing in Udacity is the final assessment. There should be a multiple choice exam on the pattern of AWS certifications that tests the concepts and how they were applied in projects and all the design choices that were made. This should be online, live, time bound, identity authenticated and should be required to get the credential. This, in my opinion, would greatly enhance the quality of graduates and of the Nanodegree program overall. AWS offers its certification credential on the basis of such exam and it is highly respected in the industry. This combined with the projects can increase a lot of respect for Udacity credential in the industry. Udacity may even consider having this exam in person in different test centers in countries through Vu Pearson's etc. 

Udacity's deep reinforcement learning course nicely blend intuition, theory and practicals of reinforcement learning.

The intuition that governs reinforcement learning can prove difficult to comprehend, but this course simplifies it and makes learning smooth. 

Also, the course starts from the very beginning methods of solving reinforcement learning problems and gradually advances to the new methods of solution used today.

High Quality and Challenging

The overall course was great! Nice and clear explanation and an outsdabding content.

The VM plataform was a bit disapointing. with some bugs that made me reset the code and redo some stuff.

The code review could be better and the assisntance was messy