DLSS? Nintendo has two patents for machine learning systems Nintendo Connect

Yesterday, Nintendo denied a Bloomberg report on 4K Switch Pro games. However, according to the free patents online, Nintendo has registered two new patents for high-end systems, which were released on the website yesterday Thursday.

According to the listings, the patents are “for learning machine image control systems and methods”, which are explicitly related to Nvidia’s AI development methods such as DLSS and Intel XeSS. These patents also claim that these methods, like Nvidia’s tensor cores, can be “accelerated by matrix multiplication hardware”. You can see the patents Here And Here.

Deep Learning Super Model (DLSS), or AI Super Resolution, is a real-time high-tech technology for computer games developed by Nvidia. The image is computed at low resolution and then measured up to set resolution using AI algorithms. With this technology, significantly more images per second should be achieved with almost the same image quality.

DLSS has been supported by various games since 2018. The new, upgraded version of DLSS 2.0 was first introduced in some games in 2020. This further development should ensure higher frame rates and be better. Both patents were filed on March 25, 2020, at the same time that Nvidia introduced DLSS 2.0. It is quite possible that Nintendo will still use DLSS technology in the future, but it worked on its own solution before the release of DLSS 2.0, especially with the Nintendo Switch.

To write this news for you now, shortly after Nintendo denied rumors that it had sent 4K Switch developer kits to developers … it’s a little weird. It is not clear how all of this relates to each other. As we mentioned yesterday: Nintendo is always working on new hardware and the next home, portable or hybrid console is already in the development field, at least on paper. So the considerations of such high-quality methods are not far off.