(I don't know if it's what they're referring to, though)
They're using deep learning as fancy Anti-Aliasing mechanism, to fill gaps which GPU didn't have time to properly raytrace or to upscale parts of images. As I understood from what I seen ML there mostly for AA.
You mean denoising? In RT raytracing you're generally short on samples and need to interpolate between them / smooth over the shot noise.
In their Siggraph presentation last week they mentioned using AI for both denoising and antialisasing (but they didn't provide anymore details about it).
It's likely mostly for denoising, I highly doubt it can have enough power to perform both denoising and antialiasing. The ray tracing presented would be sampling and then use the model to fill the gaps.
You're effectively training the network for both when training for denoising as your ground truth images are both denoised and anti aliased. With ray tracing both problems are brute forced by the same approach of throwing more samples per pixel at the problem.
There were a number of technical sessions at SIGGRAPH that went into more details, not sure if / when they'll show up on nvidia's website.