With a little help from AI, you can now create a Bob Ross-style landscape in seconds. In March, researchers from NVIDIA unveiled GauGAN, a system that uses AI to transform images scribbled onto a Microsoft Paint-like canvas into photorealistic landscapes -- just choose a label such as "water," "tree," or "mountain" the same way you'd normally choose a color, and the AI takes care of the rest. At the time, they described GauGAN as a "smart paintbrush" -- and now, they've released an online beta demo so you can try it out for yourself. The level of detail included in NVIDIA's system is remarkable. Draw a vertical line with a circle at the top using the "tree" label, for example, and the AI knows to make the bottom part the trunk and the top part the leaves.
In five lines, you can describe how your architecture looks and then you can also specify what algorithms you want to use for training. There are a lot of other systems challenges associated with actually going end to end, from data to a deployed model. The existing software solutions don't really tackle a big set of these challenges. For example, regardless of the software you're using, it takes days to weeks to train a deep learning model. There's real open challenges of how to best use parallel and distributed computing both to train a particular model and in the context of tuning hyperparameters of different models.
Nvidia chief executive Jen-Hsun Huang announced that the company has created a new chip, the Tesla P100, with 15 billion transistors for deep-learning computing. It's the biggest chip ever made, Huang said. Huang made the announcement during his keynote at the GPUTech conference in San Jose, Calif. He unveiled the chip after he said that deep learning artificial intelligence chips have already become the company's fastest-growing business. "We are changing so many things in one project," Huang said.