This is a superb lecture that creatively explains the key concepts of Decision Trees. As an added bonus, Professor Winston has a wonderful sense of humor and chooses the topic of vampires for his example tree. [Caution to those who dare to watch late at night!]
Although this is rather simple (in comparison to the other lectures) it's very practicable and valuable. ) And professor Winston is just an amazing teacher.
my guess would be... technically the tree is trying to identify some observation as belonging to some class, not necessarily taking any sort of action or making a decision... it's a bit nit-picky i guess. but i'm not sure.
the blackboard buttons are making me crazy! I can not understand the sequences that he uses to move them up and down. Would someone tell me how they work please?! Thank you :)
The two leftmost buttons are for the front one. Upper button is up, lower is down. The rightmost buttons are for the back blackboard. The big red one in the middle stops them. :)
At 43:15 I think he wants to get to the blackboard thats on the wall, so he sends up both of them at the same time. He probably hit both buttons because he doesn't know which one of them is the one that is down. But one of them is out of screen so you don't see it moving or maybe it doesn't move because it's already up.
Well, first of. I think there is only two movable blackboards on each slot, but it's a little hard to see. When he moves both of them up there is also blackboard paint behind, "on the wall" if you will, that is not going to move unless you move the wall (so there is only 2 movable blackboards for every interface). That being said the interface have 5 buttons that can be either pressed or not so that gives us 5 bits (2^5) that wen can play with (pressed or not pressed) and set the state of the blackboard in 32 different states with different combination of pressings of buttons. If we also acknowledge that a button can be in three states (selected, not pressed, pressed) then we get 3^5 so 243 different states that we can set the blackboards in. Note about the selected state: you can see that it remains green even after he released the button as the blackboard moves, indicating that the interface can actually recognize this state. It's a little unclear weather the stop button can actually be in a selected state or not. I've assumed, for simplicity, that it can.
This "dude" is a professor and has a name: Patrick Winston. iPads can be used to take hand-written notes which is the whole premise of not using phones/laptops in class/during lectures. Presumably that's why he isn't bothered by it.
39:27 Probably the main take-away - the math is just a complicated means to an end; the overarching idea of trying to identify appropriate groups/sets is fairly straightforward and intuitive.
he says its a distraction bc you can do other things during lecture. it's a good rule. I'm doing hand written notes on paper with pen... so archane! haha
Still, its not the correct format to program an artificial intelligence, this is about programming a software program. Not artificial intelligence or lifeforms. Artificial intelligences has to create these things by themselves...Learn them or "build" them
+Snorkeldykkeren And these type of tests (trees, boosts etc) are useful for the computer to "learn"...it learns by the data you input by optimizing the result.
You didn't understood the content, this is clearly not about programming a software progrram. If you build a mechanism that measures entropy by itself, it can mount the identification trees with the right order alone and very effectively .
Sorry, I still don't understand how you arrived at 4 in the shadow group. Is it because there were 4 question marks? I fully understand the other homogenous sets. Thanks.
+Marian Lenehan 3 from questionmark + 1 from no = 4 ...however that was a good discriminant for the test's quality because of the smallness of the sample size.
Fabulous teacher ! Regards from Romania !
he died last year
@Soumyo Biswas * hisses *
Give me 1000 teachers like him and i will give you 100 thousands companies like TESLA!
Feels like I'm their classmate now...!
Some of my classmates I know are, Christopher, Patrick, Krishna, Tanya, that young Turkish (I forgot her name).
Leonardo as well
0:55 - Wow, as a Romanian I was expecting it to be about gypsies. Not disappointed.
What a professor. He is really making me consider a career in teaching! Rest In Pi, Professor!
today is world pi day, the math gods have smiled on me
very underrated prof imo
This is a superb lecture that creatively explains the key concepts of Decision Trees. As an added bonus, Professor Winston has a wonderful sense of humor and chooses the topic of vampires for his example tree. [Caution to those who dare to watch late at night!]
9:44 - how split is made
16:04 - logic of splitting
31:20 equation starts
Choose features with least disorder(weighted calculation 37:00)
“30:51: We use this because it's a convenient mechanism, it seems to make sense...", what a great teacher.
His words will be quoted by next gen cyborgs on how the humans used to look down upon them 44:00 . otherwise Fabulous Teacher . Thanks
Although this is rather simple (in comparison to the other lectures) it's very practicable and valuable. ) And professor Winston is just an amazing teacher.
I just put this in my Personal Digital Assistant.
Why does he make the difference between Identification tree and decision tree?
my guess would be... technically the tree is trying to identify some observation as belonging to some class, not necessarily taking any sort of action or making a decision... it's a bit nit-picky i guess. but i'm not sure.
David Parry isn't an identification a sort of decision tho?
An identification tree is a decision tree in which each set of possible conclusions is established implicitly by a list of samples of known class
a "decision tree" is more of a holistic term identification is specific
I don't want to be a blood sucking vampire, so thanks professor. Identification trees and disorder. Now we are talking. :-)
the blackboard buttons are making me crazy! I can not understand the sequences that he uses to move them up and down. Would someone tell me how they work please?! Thank you :)
The two leftmost buttons are for the front one. Upper button is up, lower is down. The rightmost buttons are for the back blackboard. The big red one in the middle stops them. :)
At 43:15 I think he wants to get to the blackboard thats on the wall, so he sends up both of them at the same time. He probably hit both buttons because he doesn't know which one of them is the one that is down. But one of them is out of screen so you don't see it moving or maybe it doesn't move because it's already up.
Well, first of. I think there is only two movable blackboards on each slot, but it's a little hard to see. When he moves both of them up there is also blackboard paint behind, "on the wall" if you will, that is not going to move unless you move the wall (so there is only 2 movable blackboards for every interface).
That being said the interface have 5 buttons that can be either pressed or not so that gives us 5 bits (2^5) that wen can play with (pressed or not pressed) and set the state of the blackboard in 32 different states with different combination of pressings of buttons. If we also acknowledge that a button can be in three states (selected, not pressed, pressed) then we get 3^5 so 243 different states that we can set the blackboards in.
Note about the selected state: you can see that it remains green even after he released the button as the blackboard moves, indicating that the interface can actually recognize this state. It's a little unclear weather the stop button can actually be in a selected state or not. I've assumed, for simplicity, that it can.
R.I.P Patrick Winston
Rest in Peace fabulous Teacher
he passed away?
@@quangho8120 yes 19 july (source: wikipidea) feels bad he was a great teacher :(
A professor with a lot of energy.
If by "a lot of energy" you're referring to E = mc^2, then yes, I'd be inclined to agree with you. :P
Laptops aren't okay but tablets are Q_Q
Lol this dude doesn't like phones/laptops out meanwhile dude in the front is chilling on his iPad the entire time
This "dude" is a professor and has a name: Patrick Winston. iPads can be used to take hand-written notes which is the whole premise of not using phones/laptops in class/during lectures. Presumably that's why he isn't bothered by it.
at 29:48 why do we need to use L'Hospital Rule??
N/T is 0 (not tending to 0). So if we multiply anything with 0 we get 0, right??
39:27 Probably the main take-away - the math is just a complicated means to an end; the overarching idea of trying to identify appropriate groups/sets is fairly straightforward and intuitive.
31:01 any form those curves work just about the same
As we're measuring how disordered the set is
45:00 funniest joke or the lecture
The opening was beautiful! I'm so thankful for the ending of the Cold War.
My first successful Deition trees tutorial... Great man.
This is all great material, but the first 30 minutes of this lecture can really be explained in 10 minutes
What is a PDA?
when a professor stars talking about vampries ...
ok you got my atention
Why no laptops?
ruclips.net/video/Unzc731iCUY/видео.html
He explains it here
👍
Ce faceti verilor?
i have a vampire in my tv set on friday fright night.
So watered course :-( ...
Why no laptops?
he says its a distraction bc you can do other things during lecture. it's a good rule. I'm doing hand written notes on paper with pen... so archane! haha
but having a voice activated note taking system in your pocket would be okay since its not distracting
nice
Man Cristopher is getting the most attention
you mean Krishna?
Still, its not the correct format to program an artificial intelligence, this is about programming a software program. Not artificial intelligence or lifeforms. Artificial intelligences has to create these things by themselves...Learn them or "build" them
+Snorkeldykkeren i believe these techniques are used as stepping stones towards that kind of sophisticated AI. We all have to start somewhere. :)
+Snorkeldykkeren And these type of tests (trees, boosts etc) are useful for the computer to "learn"...it learns by the data you input by optimizing the result.
+Snorkeldykkeren Are you saying you are not Intelligent because you didnt create(and develop the "logic" of) your own brain by yourself
You didn't understood the content, this is clearly not about programming a software progrram.
If you build a mechanism that measures entropy by itself, it can mount the identification trees with the right order alone and very effectively .
That's Hollywood AI; this is Compsci AI.
Sorry, I still don't understand how you arrived at 4 in the shadow group. Is it because there were 4 question marks? I fully understand the other homogenous sets. Thanks.
valkyr11 Thanks!
+Marian Lenehan 3 from questionmark + 1 from no = 4 ...however that was a good discriminant for the test's quality because of the smallness of the sample size.