Can artificial intelligence increase sport performances ? The EIPACA basket example

by | May 21, 2023 | News, News

Chat GPT, Midjourney, Dall-e, Audio-gen… allow you to create text, images, sound from a few words…
These novelties intrigue, fascinate or frighten depending on who talks about them, what we know about them, and above all how we use them.

The prospects are encouraging when AI becomes a valuable aid in diagnosis or medical research, when it detects the start of a forest fire…
Opinions are more pessimistic when it puts itself at the service of the armament, or when it deprives of freedoms. The new “chatbots” integrated into social networks (My AI from tiktok, for example) create dependencies and can lead their uninformed users to believe that they are exchanging with real friends who know them better and better.
No one is able to give AI a reliable definition. The Turing test, which puts a human being in competition with an “artificial intelligence” answering questions, is one of the tools that makes it possible to perceive the effectiveness of an AI.
It would seem that this technology will soon become ubiquitous in multiple fields. Should we be wary of it, should we get started?
It seems to us that the most judicious, to be able to talk about it and decide, is to better know how it works.
As part of the basketball sports section of EIPACA, Mr. Litschgy’s students had the opportunity to immerse themselves in it, to test a very simple AI, and above all to discover its mechanisms.



 – Recognition of objects via a “learning machine” capable of distinguishing a basket, a ball, a bent or outstretched arm: It involved taking a large number of shots, then integrating this data (data) into the software who was then taught to recognize these different elements.
It was then enough to film any basket ball, arm outstretched or bent, for the application to recognize quite easily the type of object or the posture which was aimed by the lens of the camera.

Home court  is an artificial intelligence application dedicated to learning basketball. It was partly tested by the students who were able to progress in a few specific motor areas:
-Responsiveness to a signal without a balloon
-Responsiveness to a signal while dribbling
– Driving the ball (dribble) towards virtual targets
-The shots (statistics, speed)
The application is able to locate the hand of the basketball player, and awards him points when the hand hits a virtual target.
The AI recognizes the ball, and a bonus is given when the ball hits the target seen on the screen by the dribbler.
We tested it, the AI recognizes the ball, the basket, and it credits the percentage of success when it validates that the ball goes into the basket.

Discussions, surveys, debates on artificial intelligence with the main question: can AI replace a basketball coach?

Pour en savoir plus, découvrez le film qui retrace l’expérimentation réalisée au mois d’avril 2023 par les élèves de la section sortive basket de l’EIPACA :