A peek into NVIDIA’s self driving AI.

In a step toward making AI more accountable, Nvidia has developed a neural network for autonomous driving that highlights what it’s focusing on.

Sorta follow-up to my previous blog post: Is AI Mysterious?. Some of the most powerful machine-learning techniques available result in software that is almost completely opaque, even to the engineers that build it.

nvidia_car.jpg

We still don’t know exactly how AI works, the people at Nvidia still don’t know why the AI they built wasn’t following a single instruction given by its creators. Instead, it relied entirely on an algorithm that had taught itself to drive by watching a human do it.

It simply matches input from several video cameras to the behavior of a human driver and figures out for itself how it should drive. The only catch is that the system is so complex that it’s difficult to untangle how it actually works.

Nvidia is now working to open the black box by developing a way to visually highlight what the system is paying attention to. Explained in a recently published paper, the neural network architecture developed by Nvidia’s researchers is designed so that it can highlight the areas of a video picture that contribute most strongly to the behavior of the car’s deep neural network. Remarkably, the results show that the network is focusing on the edges of roads, lane markings, and parked cars—just the sort of things that a good human driver would want to pay attention to.

PilotNet.png

Nvidia’s Convolutional Neural Network architecture for Self-Driving AI.

“What’s revolutionary about this is that we never directly told the network to care about these things,” Urs Muller, Nvidia’s chief architect for self-driving cars, wrote in a blog post.

Fascinatingly enough, this compares a lot to human intelligence where we don’t actually know how to describe certain actions we do, but we just do them.

This certainly highlights the fact that in the near future we might start seeing AI those are just like us or even Superintelligent. I highly recommend reading the book Superintelligence: Paths, Dangers, Strategies by Nick Bostrom.

(Sources: Nvidia, MitTechReview, Nvidia Blog)

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s