How do RL agents really learn? | Reinforcement Learning Part-2

In this video, we present the fundamental algorithms that make Reinforcement Learning as powerful as it is today. The ideas, which originated several years ago, still find their way into today's state-of-the-art algorithms. That's why we are dedicating a video to these algorithms. Happy Learning!

============================
Do you want to learn from me?
============================

📱 Grow with us:

👍If you find this video helpful, consider giving it a thumbs up and subscribing for more educational videos on data science!

💭Share your thoughts, experiences, or questions in the comments below. I love hearing from you!

⌚Time Stamps⌚

0:00 - Intro
0:43 - MDPs
2:00 - Model
4:15 - Explore-Exploit
5:08 - Dynamic Programming
10:40 - Monte Carlo Methods
14:05 - TD Methods
15:49 - Driving Home Example
19:40 - SARSA
22:30 - Q-Learning
23:14 - Comparing SARSA and Q-Learning
24:31 - Outro

[umaraslam]

Undoubtedly the best overview of RL I found on YT. I went thru deepmind course and what not but the clarity that I got from these videos is just really amazing. Thanks for this.

[rbk.technology]

Really thank Rajthilak sir for this informative video. Please don't think bad about this message sir. Nothing personal.

But i am requesting to Nitish Sir (CampusX creator), sir you itself please create videos for the channel.

As u r the USP (unique selling proposition) in this channel.

- As the way u speak
- The way u breakdown the concept
- the best way of writing in iPad instead of slides.
These are really awesome style of teaching.

Fact:
All the people good in technology can't teach. It's a special quality.
A best surgeon need not be a Lecturer and vice versa.

So instead of anyone else. Even founder of these concepts are not needed.

We need ONLY NITISH Sir.
Please sir believing you, i took the risk of following only your playlist.

Please please kindly you yourself take videos and upload.

Last point.
Please don't worry on subscribers. It will 💯 come sir. I will share as much to clg grps and other platforms. (Also tell my friend influencers)
Your channel one day will suddenly boom with subscribers.
Please don't quit bcoz of it.

I even saw ur podcasts. Which was completely realistic. No masala added. Only plain story. Great one.

[vaibhav]

Nitish sir
Please complete machine learning interview questions as soon as possible.

[ambaradhikari]

Sir please complete gradient boosting and xgboost playlist with python implementation

[fozler]

2 videos are hidden from your reinforcement learning playlist sir. Is there any reason sir?

[arghadeepdey]

You say a lot of information in a very short span of time, please show those information in a text format in the video as well. Just like Josh Starmer's videos. It helps in letting us get used to with that info and also pay attention to what you are saying next.
Just like in 11:20, you said "Instead of expected return, they are calculated as sample return", this text could have been shown in the video as well. You skimmed through this part and then proceeded to show the equations.

[harshavardhan]

majority dimak ka upar se gaya evn after watching it fr second time, NITISH sir would have explained in a better way

[arghadeepdey]

can someone please explain, how the state values were calculated at 9:10 ?

[praphulshaw]

The video is too monotonous and I can't tell which is what, And differentiates bw one topic and another

[YadavSachin]

God but mostly subscribers are not satisfied because not a used pen .