World models, algorithms capable of generating simulated environments, represent an avant -garde of automatic learning. Today, Microsoft published a new research in Nature Detailing Muse magazine, a model capable of generating images of games and controller tickets. Unexpectedly, he was born from a training set that Microsoft built Hemorrhage edge.
Yes, like me, you had completely erased that game from your memory (or never knew that it existed first), Hemorrhage edge He is a 4 vs. fighter. 4 Developed by Ninja Theory, the best known study for his work in the Hellblade series. The ninja theory stopped updating Hemorrhage edge Less than a year after launch, but Microsoft included a clause in the game of the game that gave him permission to record games that people played online. So, if you were one of the few people who played Hemorrhage edgeCongratulations, I suppose: you helped the company do something with a commercial failure.
So what is good muse anyway? Let’s say a game designer in Blizzard wants to try an idea for a new hero in Overwatch 2. Instead of recruiting a team of programmers and artists to create code and assets that the study can eventually discard, they could use Muse to make prototype creation. The iteration is often the slowest (and expensive) part of making a video game, so it is easy to see why Microsoft would be interested in using AI to increase the process; It offers a way for the company to control fugitive development costs. This is because, according to Microsoft, Muse stands out in a world models capacity that the company calls persistence.
«Persistence refers to the ability of a model to incorporate (or ‘persist’) user modifications in generated game sequences, such as a character that is copied into a visual game,» says Katya Hofmann, senior main research manager of Microsoft Research. In other words, Muse can quickly adapt to the new game elements as they are introduced in real time. In one of the examples that Microsoft shared, you can see that the «player» character reacts immediately as two enhancers are introduced next to them. The model apparently knows that the pills are valuable and something that players would do everything possible to obtain. Then, the simulation reflects that, in the process, it creates a convincing facsimile of a real Hemorrhage edge phosphorus.
According to Fatima KardarGaming AI Corporate Vice President in Microsoft, the company is already using Muse to create a «playable real time model trained in other third -party games», and exploring how technology could help you bring old games stuck in the hardware aged to new audiences.
Microsoft says that Muse is a «first of its type» model, but that is not entirely correct. World models are not new; In fact, Muse is not even the first trained in a Microsoft game. In October, the Decart company Oasis debutedthat is capable of generating Minecraft levels. What Muse shows is how fast these models are evolving.
That said, there is a long way for this technology to leave, and Muse has some clear limitations. On the one hand, the model generates images at a resolution of 300 x 180 pixels and approximately 10 paintings per second. For now, the company is launching the pesos and data of muse samples, and a way for researchers to see what the system is.
If you buy something through a link in this article, we can win commission.
