UPDATED 18:01 EDT / SEPTEMBER 29 2016

NEWS

Nvidia demonstrates self-driving car that learned entirely from watching humans

Nvidia Corp. has released a new video showing off its latest progress in autonomous vehicle technology, this time with a self-driving car that has learned entirely from watching human drivers.

“In contrast to the usual approach to operating self-driving cars, we did not program any explicit object detection, mapping, path planning or control components into this car,” Nvidia explained in the description for the video. “Instead, the car learns on its own to create all necessary internal representations necessary to steer, simply by observing human drivers.”

In the video, Nvidia shows that its self-driving car is able to navigate through complicated road environments, including un-lined streets, construction zones, blind corners and so on. The company even noted that while the car was trained in California, it was able to drive in New Jersey without difficulty.

According to Nvidia, the car’s AI learned how to drive with data from only 20 example runs by different drivers at various times of day. “Learning to drive in these complex environments demonstrates new capabilities of deep neural networks,” Nvidia said.

Earlier this year, Nvidia wrote a blog post in which it outlined its process for training its self-driving car, which involves the use of a convolutional neural network (CNN) that taught the car how to drive. The research team trained the CNN with driving footage that was shot from a front-facing camera, which was then synced with steering data that was recorded from the drives.

“Our engineering team never explicitly trained the CNN to detect road outlines,” Nvidia said in its blog post at the time. “Instead, using the human steering wheel angles versus the road as a guide, it began to understand the rules of engagement between vehicle and road.”

Nvidia’s leap into machine learning

Nvidia has been a household name with PC gamers for some time now thanks to its high-end graphics processing units (GPUs), most notably the GeForce series, which  but the company recently made a splash in the enterprise by elbowing its way into the machine learning sector. As it turns out, those same GPUs that power bleeding edge gaming rigs are also excellent for machine learning research.

Nvidia has also on its way to becoming a major player in self-driving cars, and last month the company announced a new partnership with Chinese tech firm Baidu Inc., which would combine Baidu’s cloud-updated 3D maps with Nvidia’s hardware and machine learning capabilities.

If you want a more in-depth explanation of how Nvidia developed its self-driving AI, you can read the company’s research paper, which is titled  “End to End Learning for Self-Driving Cars.” You can also watch Nvidia’s demonstration video below:

Image courtesy of Nvidia

A message from John Furrier, co-founder of SiliconANGLE:

Your vote of support is important to us and it helps us keep the content FREE.

One click below supports our mission to provide free, deep, and relevant content.  

Join our community on YouTube

Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.

“TheCUBE is an important partner to the industry. You guys really are a part of our events and we really appreciate you coming and I know people appreciate the content you create as well” – Andy Jassy

THANK YOU