Teaching a Computer to Play Video Games at the Technion - Israel Institute of Technology

Mastering a video game requires skill, tactics and strategy. While these attributes may be acquired naturally by human players, teaching them to a computer program is a far more challenging task. In recent years, extensive research was carried out in the field of reinforcement learning and numerous algorithms were introduced, aiming to learn how to perform human tasks such as playing video games. A commonly used benchmark environment allows algorithms to train on various Atari 2600 games. However, most Atari games no longer pose a challenge to state-of-the-art algorithms. We introduce a new learning environment based on the Super Nintendo Entertainment System (SNES). The environment is expandable, allowing for more video games and consoles to be easily added to the environment. We use deep convolutional networks to train the computer to play SNES games. We show that SNES games pose a significant challenge to current algorithms due to their higher level of complexity and versatility. For some games, we are able to overcome those difficulties and suppress human performance.

Follow thins link for a paper describing this work: https://arxiv.org/abs/1611.02205

Undergraduate students: Shai Rozenberg, Nadav Bhonker
Supervisor: Itay Hubara
Signal and Image Processing Laboratory (SIPL)
Andrew & Erna Viterbi Faculty of Electrical Engineering
Follow this link to learn more SIPL: http://sipl.eelabs.technion.ac.il/

Parts of the video were taken from:
https://youtu.be/yRsjJ-u2ipw
https://youtu.be/4Jc51w_VCzE
Music by:
http://www.purple-planet.com/

Видео Teaching a Computer to Play Video Games at the Technion - Israel Institute of Technology канала SIPL Technion