Professional Services
Order outsourcing

How Ubisoft Uses AI & Machine Learning to Build Games

During the Ubisoft Developers Conference in Montreal last week, Ubisoft shared a closer look at the way they use AI and machine learning.

During the Ubisoft Developers Conference in Montreal last week, Ubisoft shared a closer look at the way they use AI and machine learning. The team, for example, used smart algorithms to teach NPCs how to drive in Watch Dogs and create bots in For Honor to playtest combat mechanics.

The most important part, though, is hidden in “the Minority Report of programming.” The studio’s Commit Assistant is the project of La Forge, the Ubisoft Montreal collaborative R&D division, and it is something truly amazing. 

When a programmer writes a new line of code, Commit Assistant searches for parts of the code that are adding bugs. The algorithm is based on a decade’s worth of code changes and bug reports from Ubisoft games at every stage of development. The project is said to analyze bugs and create a “signature” that estimates the impact of the new code and the probability of future problems.

Ubisoft states it finds 6 out of 10 bugs accurately, and presents a 30% false alarm rate, and the things is that the tool re-trains itself. Commit Assistant also gives information about potential issues and suggests a fix for it. 

As of this moment, the project saves programmers 20 per cent of time. Ubisoft’s Yves Jacquier states it is an estimate based on a recent test case:

To make sure it worked before we put it in the hands of production, we froze the process and observed a couple of productions for three to six months, I believe. Just to see how they behave, what are the bugs they created. Based on their history, we simply ran a program saying how many of those bugs would we have caught in that time. By doing that and estimating false alarms, etc. we came to the assumption that it was about 20 per cent of the work that had been done by this team during this time.

When we first introduced mocap, it was perceived as a threat, like it would cut animators’ jobs. It was the opposite. The mocap was facilitating animation a lot. It was helping animators to focus on where they have real added value, which is the emotion, the spectacular aspect of the animation. We created more spectacular animation, we wanted more animation in games, which created more animators at the end of the day.

Yves Jacquier

Just think about the possibilities of these technologies: Ubisoft’s AI programs can now generate assets for games that are about 80 per cent as good as what artist can do.

Today the AI can redo the job for very specific things, such as text-to-speech, as long as you don’t put in emotion. Or animation, but only for locomotion. Facial animation, as long as you don’t expect emotion on the face but mostly leaps, moving, etc.

I think it will take three years before you can have a fully generated NPC in a game,” he said. “A character with its own animation, modeling, facial expression, text-to-speech with different emotions or variation. And that’s totally comparable to something you’re doing today with animators, people who are specialists in speech, modelers, etc.

Yves Jacquier 

You can learn more about the Commit Assistant in a thorough report by gamesindustry.biz.

Join discussion

Comments 2

  • rlawsfamily

    Its great of Ubisoft to share so much information with the industry.

    There will be one thing however which can never be automated - the creative spark and genius that leads to unique and thrilling gameplay. Game designers will become more crucial than ever before, with the rest becoming more like hygiene factors.



    ·6 years ago·
  • Nexus-8

    In 2070, vast pleasure domes acomodate and entertain millions with the latest video games completly produced by AI, while your personal replicant pleasure model satiates your deepest and darkest fantasies.



    ·6 years ago·

You might also like

We need your consent

We use cookies on this website to make your browsing experience better. By using the site you agree to our use of cookies.Learn more