New Unity graduate fellowship program will research AI and machine learning
Initiative will support grads exploring the possibilities for content creation and NPCs that learn new behaviour
Unity has announced a new fellowship program that seeks to identify the potential for smarter artificial intelligence in games.
The venture was announced during this week's Unite Europe conference in Amsterdam, and gives two graduate research fellows the chance to spend six months working on projects investigating the cutting edge of machine learning algorithms.
That might not sound overly sexy, but it translates to some interesting implications for games development in the future. For example, machine learning and more advanced AI could lead to non-player characters that show new behaviour as a game progresses, rather than purely adapting to what the player is doing.
It could also be used to create new animations, graphics and dialogue for games in a way that saves time for the developers but still creates a game assets that are unique.
The graduate fellowship program is a joint venture between Unity Labs and Unity's AI & Machine Learning Group, and applications are now open. The deadline is September 9th, but at the time of writing the program's website has yet to go live. Unity has promised more updates will be issued via its Twitter and Facebook feeds.
In an official blog post, the firm's VP of machine learning Danny Lange said: "We recognize that schools around the world, offer programs to teach technical skills in computer games programs in computer science or related fields. We want to provide enrolled graduate students an opportunity to go beyond, and work on relevant and applied research related to games, and enable them to bring this research back to their schools and communities of practice."
AI is rapidly becoming a hot topic among game developers as they seek new ways to make virtual worlds more immersive, believable and reactive to the player. Earlier this month at E3, Electronic Arts announced a new R&D division called SEED, which will be exploring something similar through research into neural networks and deep learning.