Tyler Umansky

27:55

yes

Jack Newsom

27:55

yep

Jonathan Kim

27:56

yes

Joshua Wu

27:57

yes

Luke Cheng

27:57

Ok for me

ssingh17

27:58

yes

Chenyue Cai

27:59

yes

Calvin Wong

28:00

y

Estea

31:52

could you go back to utility quickly?

Estea

32:01

just the last part of last slide

Jonathan Kim

32:26

Slides for today: https://drive.google.com/file/d/1fxTp0EeyDJizuDYc7jjHDRs0eXQaGqAC/view?usp=sharing

Jonathan Kim

32:32

this is the pdf for the slides

Estea

32:41

ty!

Amit Palekar

41:23

The states “utility” is the score associated with that state right?

Joshua Wu

43:04

we technically didnt need to explore the 4 and 6 nodes correct?

Calvin Wong

43:21

alpha beta pruning yee

Amit Palekar

43:25

does this algorithm depend on the fact that the opponent plays optimally

Estea

43:30

can you go over the successor functions?

Julius Tereck

44:34

But if you try to exploit imperfect play, doesn't that open you up to being exploited as well?

Andrew

45:02

yea but you're assuming they aren't good enough to play optimally

Danny Sallurday

45:28

I suppose there would be an extra component of the algorithm to weigh risk/reward against suboptimal opponents

Zackoric

46:31

I thought the ghosts were master minds

Rafael Flores

46:36

gopacman

Mengying Ju

46:54

What

Mengying Ju

47:09

What's the goal test? Eat all the dots? Eat a ghost?

Yewen Zhou

47:19

if the ghost is moving randomly, would the action of pacman be deterministic?

Abayjeet Singh

48:11

I think yeah based on movement of ghosts

Chenyue Cai

48:35

In the case when ghosts are not acting optimally, are we introducing probability in the evaluation?

Vishal Raman

49:10

You could compute expected values of moves to make a choice in that case

Julius Tereck

49:13

I think the policy would be deterministic, but the actual action would random, and depend on the movement of the ghost

Yewen Zhou

51:30

but since the movement of ghosts are random, how would the pacman act based on the actions of ghosts? By introducing probability?

Julius Tereck

52:59

Yeah I think then the policy would be to take the action with the best expected value

Abayjeet Singh

53:24

is it possible for movement of ghosts to complete random? it will be like pseudo random

Calvin Wong

53:55

That's the question of coming up with a function that generates random numbers basically, which is really hard

Calvin Wong

54:03

because how do you define random

Calvin Wong

55:26

but pseudo random yes, depending on what you mean by "random"

Julius Tereck

01:00:11

So how would you do that child ordering?

Carlos Calderon

01:00:37

can you clarify how the ordering improves the algorithm?

Chenyue Cai

01:01:29

why could the intermediate node be wrong when we use <=?

Abayjeet Singh

01:01:54

Can u give an example where ordering improves algo?

Julius Tereck

01:03:13

So if this is only for zero-sum games, where does it fail for non-zero-sum?

Estea

01:03:56

was the v <= 10 the wrong node

Estea

01:04:12

and 0 would've been right

Estea

01:04:13

?

Abayjeet Singh

01:04:58

Got it!!

Jonathan Kim

01:05:02

what if the value of 5 was 1?

Julius Tereck

01:06:19

Ahh okay. So they would have to maximize each others utilities instead of minimizing their own

Calvin Wong

01:06:24

I think ye Estea, but what he was getting at was if there was a smaller value on the right (eg: -1) then we still get 10 as the minimax root value

Jonathan Kim

01:06:26

scratch my question

Calvin Wong

01:06:42

i think

Estea

01:07:02

So intermediate nodes are wrong; wouldn't that make the root node also wrong?

Calvin Wong

01:07:03

but then the true value should've been -1 rathre than 0

Marco Gellecanao

01:07:03

Can you please explain again why the time complexity for minimax with pruning is O(b^(m/2))

Marco Gellecanao

01:07:29

oh ok nevermind then thanks

Calvin Wong

01:07:42

no since we pick the 10 either way

Danny Sallurday

01:07:44

Intermediate nodes would be wrong if they have a value already worse than the root node; the root node stays optimal

Danny Sallurday

01:07:53

Aka right

Estea

01:08:02

k..

Calvin Wong

01:08:10

^ Danny bump

Calvin Wong

01:09:35

So like no matter if we picked the 0 or the -1 as the right value, the top node maximizes so it would always choose the 10 @Estea

Joshua Wu

01:09:53

f

Estea

01:09:54

I see.. hence the pruning?

Danny Sallurday

01:10:02

yes

Estea

01:10:32

Ty @ Calvin & Danny!

Calvin Wong

01:10:44

Yee

Tyler Nunez

01:10:45

F

Calvin Wong

01:10:58

lol f in the chat

Chenyue Cai

01:20:02

are -2 and 9 heuristics? How do we know value if we don’t reach the terminal

Calvin Wong

01:20:38

Yep heuristic estimates

Estea

01:22:08

can't we limit actions using cost

Estea

01:23:19

k..

Yewen Zhou

01:25:25

in this case what if on the board both sides don't have queens anymore? Then the heuristic will stay at 0 right?

Amit Palekar

01:25:33

wait how did he fix the thrashing

Danny Sallurday

01:27:42

So what separates evaluation functions from heuristics ?

Yewen Zhou

01:28:03

how do we know the ghosts are DFS not BFS?

Carlos Calderon

01:28:17

what do you mean “the actual terminal utility comes into play”?

Calvin Wong

01:28:39

Think that means they find the actual value rather than heuristic

Calvin Wong

01:29:05

Fixed the thrashing with a better heuristic

Estea

01:31:03

better heuristic not evaluation function?

Ferdie Taruc

01:31:30

^ the evaluation of the utilities are based on proximity to the dots now, rather than seeing just the dots exist when depth = 2

Calvin Wong

01:31:33

from my understanding they're the same?

Ferdie Taruc

01:31:48

its pretty much the same, you use different heuristics to determine an evaluation function I think

Calvin Wong

01:31:48

not too sure

Estea

01:32:01

i see.. so they r similar

Tyler Umansky

01:33:26

Are lecture recordings posted somewhere yet?