MONTREAL – A partnership between the federal agency NSERC and the interactive entertainment product producer Ubisoft has resulted in the formation of a research chair on artificial intelligence at the Université de Montréal.
The chair, named NSERC-Ubisoft Industrial Chair on Learning Representations for Immersive Video Games, will focus on three activities – the collection of data for the creation of an improved, immersive experience customized in real-time, thanks to the generation of varied automated elements with the adaptable artificial intelligence; the development of techniques to better equip artists and 3D animators in the creation of high-quality, immersive environments, namely the displaying of rich and flowing video content that takes into account the limitations of game consoles; and the sharing of knowledge, experience and tools between academics and technologists, in order to stimulate creativity and innovation with the purpose of increasing the productivity of production teams.
About 30 students and researchers will work within Ubisoft production teams to bring together video game engineers and developers’ know-how with the research methods of professors and students at Université de Montréal. The research, whose main goal is to further develop learning algorithms known as deep algorithms, should result in significant advancements in the fields of machine learning and artificial intelligence (AI).
By analyzing data on the player’s in-game behaviour, it will be possible, for instance, to offer him or her a new dimension of interactivity with the help of AI. The AI will "learn" from the data and "adapt" the content of the games and the recommendations made to the player in accordance with his or her behaviour.
"Our teams of professors and students have been collaborating for more than two years to advance research on data treatment for the creation of learning algorithms that modify in-game experience", said Professor and chairholder Yoshua Bengio, in Wednesday’s announcement. "The amount of technological data to process is always increasing and becoming harder for developers to deal with. Research work will not only help the quality and adaptation of content, but also the efficiency and productivity of developers responsible for the creation of the content."
