Hardest BOT Competitor Ever? Meet MIT’s AI Philip That Beats Human Gamers Like In Nintendo's ‘Super Smash Bros’

• comments
• 
• 

Every gamer has their best score. Unfortunately, MIT introduces an artificial intelligence that is integrated with a system powerful enough to beat any high score. Yet, if it doesn’t hold, it would rather commit suicide than to lose.

According to Daily Mail, MIT had developed AI Philip in their very own Computer Science and Artificial Intelligence Laboratory (CSAIL.) A video was released wherein Nintendo's very own ‘Super Smash Bros’ was played by a top ranking player was beaten by the video game AI.

The method to develop AI Philip was reported to have a neural network model to be trained. In which the training process includes feeding it coordinates of objects in the game which are said to be players and ledges. The video game AI was also said to be given incentives to be sure that it would win.

BusinessInsider then noted that the study led by a grad student at MIT studying artificial intelligence, Vlad Firoiu was published in ArXiv. It was then reported that AI Philip also had learned Q-Learning and Actor-Critic which are well-examined algorithms which enabled it to play “Super Smash Bros.”

Furthermore, in the video playing along is identified to be the 50th best ranking player in the world, Mafia, also known as James Lauerman. "It was hard to discern any real trial or error it was doing, let alone what strategies it was applying," he said. AI Philip is then discerned to not play like a human though since it has flawless fast moves which lead to its edge in the play as reported

After the study success on Nintendo’s ‘Super Smash Bros,’ Firoiu mentioned that he would further continue the study. He then concluded that the next study would have gameplay with human-level reaction time. In which is said to eliminate bots unexpected strategies and place it in a more relatable human move.

For more studies like Gambling Addiction Problems wherein slot machines are said to put gamblers in a trance, visit us here.