Seattle Sports Center – The latest chatter on Seattle sports teams.
Artificial Intelligence plays football (with commentary)

Artificial Intelligence plays football (with commentary)

July 24, 2017



here is the original video published 4 years ago without the commentary:

This is how my robot learns how to play a new sport. First, the robot has to learn the rules, goals, and procedures of the sport. He can accomplish this by reading books or attending school classes, like physical education.

Once the knowledge is stored in the robot’s brain, the second step is to store linear pathways in memory regarding the sport. The only way to do this is to practice playing the game. While the robot is practicing the sport, he is using logic to determine where the rules, goals, and procedures should be stored in these linear pathways.

After weeks of playing that sport, the robot creates a self-created decision tree in memory regarding the sport. This decision tree is almost like a situation based system, that uses nested if-then statements. At the beginning of the game, you do this. at the ending of the game, you do that. in this situation, you do this. in this situation, you do that.

In this example, the decision tree belongs to the game of baseball. All the pathways are stored in a timeline. For instance, at the beginning of a game, the robot has to follow orders from a coach. at the ending of the game, the robot has to shake hands with the opponents. if the robot played as the pitcher, he has to follow these rules and strategies. if the robot played as the left fielder, he has to follow these rules and strategies. Thus, this decision tree formulates a self-created, situation based system, to play baseball.
The more the robot practices the sport, the more optimal the decision trees are to play baseball. New strategies are discovered, and if they are good ones, will replace old strategies from memory. This type of re-enforcement learning will optimize the pathways in the decision trees.

the decision trees eventually turn into a self-created computer program to play baseball.

This is how my robot learns all sports, ranging from football to tennis to golf. In fact, he can use this learning method to learn every single human task. He can use this method to learn to drive a car or fly an airplane or cook in a restaurant. This type of learning is totally different from machine learning because there are no training involved. The robot learns a new skill just like a human being, by reading books and through repeated practice.

This learning method can be used to learn any human job.

the data structure to Human-Level AI:

website:
http://www.humanlevelartificialintelligence.com

tags:
call of duty, ai plays call of duty, human level artificial intelligence, ai, artificial intelligence, artificial general intelligence, true ai, strong ai, human level ai, cognitive science, ai plays video game, robot plays video game, agi, digital human brain, human intelligence, human brain, human mind, human thought, ai plays role playing games, ai play rpg.

source

Leave a Reply