This paper reviews different applications of machine learning techniques in a development of AI player for a collectible card video game Hearthstone: Heroes of Warcraft. Its main objective is to present selected aspects of the game, in which machine learning can help an artificial player to make better decisions and increase its winning chances. Moreover, the paper describes some of experiments conducted by our research team at Silver Bullet Solutions, as a part of the research on GRAIL – a general framework for designing AI players in video games.

You can read whole article here.