What is the Minimax Algorithm? - Artificial Intelligence
ฝัง
- เผยแพร่เมื่อ 7 มิ.ย. 2024
- The minimax algorithm is one of the oldest artificial intelligence algorithms ever. It uses a simple zero sum rule to find which player will win from a current position.
This is arguably the most powerful and basic tool for building game playing artificial intelligence. It captures the intentions of two player games fighting against one another. Minimax is the foundation for algorithms like alpha-beta pruning and iterative deepening.
Engines like deep blue use optimized versions of minimax to play chess, while other computer programs have also been successful at playing games with minimax.
Source Code (Complicated, glance over the evaluate function to get a gist):
github.com/gkcs/ChainReaction...
00:00 Basic Game Tree
01:17 Basic Algorithm - Winning
03:42 Heuristics
06:40 Benefits of Heuristics
07:34 Thank you!
#minimax #minmax #game-tree
This video is really great for getting across the intuition, thank you!
Clear and helpful explanation. Thanks
Very well explained.. Thanks !
Really easy to understand thank you so much brother :
Great explanation, made it easy for me to understand :)
Thanks!
Great video! Thanks a ton!!
Thanks!
hey man!....thanx for explaining.....May i knw how did you get the values for the leaf node?
loved the heuristic example :)
Perfect, thank you!
Thanks a bundle!
really great explaination, thank you sir!
Would it be possible to do a breadth-first search on the minimax tree instead of depth-first/iterative deepening? Or maybe iterative deepening can be done without visiting the same nodes so many times?
Amazing!
Thanks a lot! :D
Thanks!
Totally made the concepts clear , Thanlyou !!
Thanks!
easy and effective!
0:40 you say positive infinity is a win for white and negative infinity is a loss for black. Shouldn't it be that negative infinity is a win for black because it is trying to get the lowest score?
I messed up 😅, thanks for pointing that out!
That’s racist!
I have a question , how the heuristic function work ? 😞
even i was thinking this
@@johnmadsen37 really man
Fantastic, explained better than any MIT, berkley video.
Thanks Jamie :-)
Thats cool, i am from Britain, you need to move here, we are crying out for good teachers like you, who can mix complexity with intuition. I only speak my native language, but you speak it far better than myself , very clear and concise.
I came here from that video too. I would not say either of the videos is better or worse than the other, they just use different kind of interpretation. Both of them helped me to understand the topic deeper, so thanks for the video. :)
0:38 *Win for Black
Thank you
How would this be applied to a fighting game do you know?
Hi gaurav I am having a small doubt that how we handle the case if tree is not complete
Hey Tarun, you can return the max/min among all the nodes with the same parent. For a single node you return the value itself. For any other scenario, you choose a win if any of the remaining nodes is a win for you. Else you choose a draw if any of the nodes is a draw. Else you have to choose a loss. So you are essentially using a Max(all child node results) to try to win.
great vid:)
that's so awesome
It is 😁
Easy and simple ❤️
What tool you are using for presentation?
Google Slides! 😁
Awesome video !!
Could you please also attach a link to a problem (beginner-friendly one :P) related to this topic ???
Hi Baidhar!
Thanks for the feedback!
Here is a tic-tac-toe game you can use minimax on:
www.hackerearth.com/practice/algorithms/string-algorithm/manachars-algorithm/practice-problems/multiplayer/tic-tac-toe/
Thanks a ton !!
good video
Better explanation than MIT.
Haha. I am honored overwhemingly :-)
The profs there are awesome though.
Which software did you use to draw and explain?
Google Slides 😁
@@gkcs Well that was quick. And btw I am a great admirer of your way of teaching. Really appreciate your efforts to make difficult concepts look easy. :)
Thanks!
how to apply the minimax algorithm to the card game "Beat the Buzzard"?
Hey utseb1,
Please describe the question in detail here.
i want to find the best strategy for this game www.funagain.com/control/product?product_id=012770&reviews_show_all=Y and I want to apply the minimax algorithm but I have no idea how because it is an imperfect information game
I understand because of this
tomorrow is my exam and you are clearing my doubts in 5 mins which my bozo prof was not able to do in 6 months 😇
I was listening very seriously til hagrid came on screen!
Oh we have aragog on the book too!
which site are you using to present this? Can anyone tel?
Google slides :)
@@gkcs Thanks
Why this is so ez with you but not with ma proffesor. He was shittalking about it in 30 minutes but you explained it in on video and I am not even native english speaker!
When in doubt follow the spiders
Hahaha! Heuristic right there :D
toward the beginning, "positive infinity would be a win for white, and negative infinity would be a loss for black" means white always wins!
Damn, I should have thought that through :p
What is R & B...?
Red and Blue.
positive infinity would be win for white and negative infinity would be loss for black :D
Yup!
Gaurav Sen loss for black is the same thing as win for white isn’t it?
Ya, is should have been win for black
Thank you thank you thank you sooo much dada....by the way are you junior from me or senior?I am now 27+.So can I call you by your name or I call you as dada?
Gaurav is fine 😁
@@gkcs Ok so gaurav you and me we are off same age right?will you be my friend gaurav?facebook e aachho tumi?are you in facebook?
The diction is too fast and badly delivered. Couldn't understand much of it.
Why can't you reduce the speed and watch the video till you understand!
Positive infinity win for white
Negative infinity loss for black
In both cases black is losing.
Yeah I messed that up 😅
Negative infinity is a loss for white
No bashing or anything , but really badly illustrated