On Monday, March 18, Nvidia launched A.I. Playground, a new online space where anyone can experience all the raw power of deep learning research. The new website currently features three demos for image editing, styling, as well as photorealistic image synthesis.
As a part of this new online space, you can upload your own image and edit photos and experience Nvidia’s “Imagine InPainting” research project. The technology uses the power of artificial intelligence and can edit your images by removing certain content and filling in the resulting holes. Researchers used Nvidia Tesla V100 GPUs as part of their original work on this project. This was originally published last spring at the International Conference on Learning Representations in New Orleans.
A second demo in the online space allows you to demo Nvidia’s “Artistic Style Transfer” project. This experience takes the aesthetics of a primary uploaded image and can apply it to a second image with realistic results. For instance, you can upload a photo of yourself and have it returned to you painted in the style of a Vincent van Gogh piece. Originally announced in August 2018, researchers behind this project leveraged the power of Nvidia Titan XP GPUs. Researchers also had to train a convolutional neural network on 80,000 images of people, scenery, animals and moving objects as part of this research project.
The final available demo is of Nvidia’s photorealistic image synthesis model. This uses conditional generative adversarial networks to create photorealistic images and virtual environments. Using Nvidia’s Quadro GPUs, the technology behind this system was originally disclosed in June 2018 at the annual Computer Vision and Pattern Recognition Conference in Salt Lake City.
There are plans to add more demos in the future, and Nvidia believes this new experience is for people with a wide range of interests, including photo editing. “Research papers have new ideas in them and are really cool, but they’re directed at specialized audiences. We’re trying to make our research more accessible. The A.I. Playground allows everyone to interact with our research and have fun with it,” Bryan Catanzaro, vice president of applied deep learning research at Nvidia, said in a statement.
Nvidia’s GPU Technology Conference is underway in San Jose, California, and you can expect to hear more about artificial intelligence, gaming, cloud services, science, robotics, data centers, and deep learning throughout the four-day event.
