AI Playing Games: Crash Course AI #12
video description
Date: 2022-04-04
Related videos
Comments and reviews: 5
Grand
I'm still waiting for a AI that could play a competitiv game of Civilazation against human players whiteout massiv cheating. Because it does feel like it should be possible but on the other hand it's so extremely complex. And I think the biggest issue would be about that there is no clear account on a particular good or bad move. You simply can't evaluate the impact a certain choice had on the game in a hole. Maybe you could analyse lots of data and winning strategies and maybe you can teach the AI to hit a snowball effect early on and take home a win, but I hardly doubt that there will be an AI fit for human learning and counter strategies. At least when it comes to conspiracy an betrayal AI isn't fit for challenging humans yet.
reply
I'm still waiting for a AI that could play a competitiv game of Civilazation against human players whiteout massiv cheating. Because it does feel like it should be possible but on the other hand it's so extremely complex. And I think the biggest issue would be about that there is no clear account on a particular good or bad move. You simply can't evaluate the impact a certain choice had on the game in a hole. Maybe you could analyse lots of data and winning strategies and maybe you can teach the AI to hit a snowball effect early on and take home a win, but I hardly doubt that there will be an AI fit for human learning and counter strategies. At least when it comes to conspiracy an betrayal AI isn't fit for challenging humans yet.
reply
Lord
An interesting video but just one point about the number of possible games of tick tack toe. You can reduce the number significantly by using symmetry. The first move for example can be reduced down from 9 moves to just 3. The centre, the middle of any row/column and any corner. All other moves can be created using rotation and/or reflection.
reply
An interesting video but just one point about the number of possible games of tick tack toe. You can reduce the number significantly by using symmetry. The first move for example can be reduced down from 9 moves to just 3. The centre, the middle of any row/column and any corner. All other moves can be created using rotation and/or reflection.
reply
rchave
Cards Against Humanity. The AI will need to understand innuendo, have a very flexible use of language, and intuition as to what is taboo, physically implausible, dark, or filthy. That's the greatest challenge.
reply
Cards Against Humanity. The AI will need to understand innuendo, have a very flexible use of language, and intuition as to what is taboo, physically implausible, dark, or filthy. That's the greatest challenge.
reply
PennyAfNorberg
in the summer of 96 I took a curse in neural network ( using a second order training algorithm) and me and my group tried to make a connect -5 (infinite grid. it was fun.
reply
in the summer of 96 I took a curse in neural network ( using a second order training algorithm) and me and my group tried to make a connect -5 (infinite grid. it was fun.
reply
Александр
As far as i know there was no official -best human vs Google AI- match, there was one unofficial though, but technically AI didnt beat humans in Starcraft yet: )
reply
As far as i know there was no official -best human vs Google AI- match, there was one unofficial though, but technically AI didnt beat humans in Starcraft yet: )
reply
Add a review, comment
Other channel videos
