Markov Decision Processes Three - Georgia Tech - Machine Learning

  Рет қаралды 52,574

Udacity

Udacity

Күн бұрын

Пікірлер: 20
@demidevil666
@demidevil666 5 жыл бұрын
This discussion helped me tremendously while writing my Bachelor's thesis. I love the fact that I'm listening to a vivid conversation and not a dry lecture.
@hassanfarooq7379
@hassanfarooq7379 5 жыл бұрын
Its best way to learn when there is a funny student there like Michael. Great work!!!
@Emperorerror
@Emperorerror 5 жыл бұрын
Could not disagree more with the others here. Having the student helps a lot
@misterproject8
@misterproject8 4 жыл бұрын
Agreed!
@muhammadrehan4752
@muhammadrehan4752 6 жыл бұрын
Its a fun way learning. Thumbs up guys. Love from Pakistan :)
@misterproject8
@misterproject8 4 жыл бұрын
In a sense, we do actually just use the present to make decisions. We don't know for certain what happened in the past, we only trust that our memories and records are good enough to represent them.
@huesOfEverything
@huesOfEverything 2 жыл бұрын
And they had to pass it to me one generation at a TIME because you know - Markov Property :D This guy is just hilarious
@maba-n3j
@maba-n3j 4 жыл бұрын
Hi, I am from Russia and I really appreciate how you mentioned Russia! It was really funny))) Thank you for the video!
@imalive404
@imalive404 8 жыл бұрын
@1:52 the professor tells that the obvious problem with that is if you have to remember everything from beginning of time you are only going to see every state once. Could you please explain it ?? :)
@_SimpleSam
@_SimpleSam 8 жыл бұрын
+imalive The suggestion is that you remember past states by including them in the current state. Meaning, you would have a super long 'id', a lot of data, to define the current state. The problem with that is the 'state' gets big very quickly, and more importantly, each new state is necessarily unique and 'new' given that the history is infinite and the state definition growing. If that makes any sense :P
@imalive404
@imalive404 8 жыл бұрын
SimpleSam Thanks man :)
@_SimpleSam
@_SimpleSam 8 жыл бұрын
imalive Figured you moved on, but I saw it and figured I'd try to help the next digital wanderer :)
@stanley_george
@stanley_george 4 жыл бұрын
If some day I end being a lecturer of comp Sci.. I would definitely use these funny lines about Russians and historians
@DBBBB
@DBBBB 6 жыл бұрын
Am I the only one who found those jokes incredibly funny?
@hassanfarooq7379
@hassanfarooq7379 5 жыл бұрын
haha, this is the funniest yet best way to learn.
@stefantincescu7169
@stefantincescu7169 5 жыл бұрын
You waste a lot of time talking to that guy.
@civilowl965
@civilowl965 6 жыл бұрын
Great content but could you get rid of the other guy. He's really annoying and unfunny and unnecessary. Thanks.
Markov Decision Processes
43:18
Bert Huang
Рет қаралды 78 М.
Вопрос Ребром - Джиган
43:52
Gazgolder
Рет қаралды 3,8 МЛН
Жездуха 41-серия
36:26
Million Show
Рет қаралды 5 МЛН
Какой я клей? | CLEX #shorts
0:59
CLEX
Рет қаралды 1,9 МЛН
Markov Decision Processes - Computerphile
17:42
Computerphile
Рет қаралды 183 М.
Markov Chains Clearly Explained! Part - 1
9:24
Normalized Nerd
Рет қаралды 1,3 МЛН
Policy and Value Iteration
16:39
CIS 522 - Deep Learning
Рет қаралды 162 М.
Markov Decision Process (MDP) - 5 Minutes with Cyrill
3:36
Cyrill Stachniss
Рет қаралды 35 М.
Solving MDPs
15:32
CIS 522 - Deep Learning
Рет қаралды 14 М.
Вопрос Ребром - Джиган
43:52
Gazgolder
Рет қаралды 3,8 МЛН