Unity introduces Machine Learning Agents to help AI advancement
New feature goes into beta, primarily for researchers but engine firm keen to see what developers can do
Unity has put a new feature into open beta that ultimately aims to help drive forward research into more comprehensive artificial intelligence in video games.
The premise of machine learning is that, rather having to manually code the behaviour for elements such as enemies and NPCs, characters learn from their interactions and the virtual environment around them and adapt their behaviour going forward.
To that end, Unity has introduced Machine Learning Agents into open beta today. The SDK for ML-Agents is open source and comes with a set of example projects and baseline algorithms to get developers started.
"This release is only the beginning, and we plan to iterate quickly and provide additional features for both those of you who are interested in Unity as a platform for Machine Learning research, and those of you who are focused on the potential of Machine Learning in game development," the firm's deep learning and AI researcher Arthur Juliani writes in an official blog post.
"While this beta release is more focused on the former group, we will be increasingly providing support for the latter use-case. "
The blog post goes into much more detail on how developers can get started with Machine Learning Agents, and how they can be used. Back in June, Unity announced a new fellowship program that would support graduates exploring the possibilities for machine learning and deeper AI.
Unity is not the only interested in how machine learning can advance AI in games. In addition to start-ups like Spirit AI, major publishers like Electronic Arts are exploring the potential for such tech. At E3 2107, EA announced it has set up a new experimental division, SEED, dedicated to this and hopes to introduce stronger AI into its games within five years.