UPDATED 17:54 EST / DECEMBER 04 2019

photo-in-aws-re-invent-2019-set-2-las-vegas-nv-day-1-google-photos AI

Fight to finish: AI-on-AI and dev-on-dev speed up model training

Why would anyone trust artificial intelligence? It’s basically the same reason we trust ourselves to make decisions every day. We examine data, reference historical outcomes, and project into the future. The speed and number of cycles from reference to prediction — and all the data throughout — determine just how good AI can get.

“We trust humans to do certain things. … There’s a system where [a] neurosurgeon probably had to go through a ton of training, be tested over and over again, and now we trust that he or she is doing the right thing,” said Naveen Rao (pictured), vice president and general manager of the artificial intelligence products group at Intel Corp.

Trust in AI can come through analytical analysis or empirical evidence like in the case above, Rao pointed out. Whether we know the science behind AI or not, we’ll want to see it perform reliably many times over before we send it out into the wild.

Rao spoke with  Dave Vellante (@dvellante), host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, and guest host Justin Warren (@jpwarren), chief analyst at PivotNine Pty Ltd., during the AWS re:Invent event in Las Vegas. They discussed how humans and technology are helping to speed AI to greater performance and reliability. (* Disclosure below.)

Parallel legs kick AI ball forward

Anything that reduces the length of model-building and training cycles helps developers. There is a kind of AI-on-AI approach called Generative Adversarial Networks in which two neural networks constantly work against each other. For example, one side is creating a fake picture while the other tries to guess if it’s fake.

“Once it can’t tell anymore, you’ve kind of built something that’s really good,” Rao said. 

Intel partners with Amazon Web Services Inc. to combine its tech — like AI-enhanced chips and its OpenVINO AI toolkit — with AWS’ offerings like the DeepRacer 1/18th-scale car powered by reinforcement learning. There is an ecosystem of developers constantly tinkering with the tech, trying to move the ball forward, according to Rao. There are competitions where developers come together to demonstrate what they’ve accomplished. They showcase how they’ve trained the Intel chip in the car for better driving, etc.

“I think competition is how you push technology forward,” Rao concluded. 

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s coverage of the AWS re:Invent event. (* Disclosure: Intel Corp. sponsored this segment of theCUBE. Neither Intel nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE

Since you’re here …

Show your support for our mission with our one-click subscription to our YouTube channel (below). The more subscribers we have, the more YouTube will suggest relevant enterprise and emerging technology content to you. Thanks!

Support our mission:    >>>>>>  SUBSCRIBE NOW >>>>>>  to our YouTube channel.

… We’d also like to tell you about our mission and how you can help us fulfill it. SiliconANGLE Media Inc.’s business model is based on the intrinsic value of the content, not advertising. Unlike many online publications, we don’t have a paywall or run banner advertising, because we want to keep our journalism open, without influence or the need to chase traffic.The journalism, reporting and commentary on SiliconANGLE — along with live, unscripted video from our Silicon Valley studio and globe-trotting video teams at theCUBE — take a lot of hard work, time and money. Keeping the quality high requires the support of sponsors who are aligned with our vision of ad-free journalism content.

If you like the reporting, video interviews and other ad-free content here, please take a moment to check out a sample of the video content supported by our sponsors, tweet your support, and keep coming back to SiliconANGLE.