2-on-2 A.I. football players trained with reinforcement learning M.L. playing against each other

 στο Programming  επισημασμένο με ετικέτα a.i. / ai / artificial intelligence / football / m.l. / machine learning / ml / reinforcement learning / rl / unity Από

This lengthy video shows several hours of 4 artificial intelligence agents playing football. The players were trained with machine learning. This means that they were never instructed on how to play football or what their purpose was. After several tries, they found out that getting the ball into the opponents’ goal rewards them, while if the opponent achieves the opposite effect, they receive a penalty.

This video was produced using the Unity game engine.

This post is also available in: Αγγλικα

Απάντηση

Αυτός ο ιστότοπος χρησιμοποιεί το Akismet για να μειώσει τα ανεπιθύμητα σχόλια. Μάθετε πώς υφίστανται επεξεργασία τα δεδομένα των σχολίων σας.