I remember some years back I was sitting with some friends watching NASCAR. Being in marketing, everything I watch, read or see on the web I always notice the marketing aspects. I would sit there watching those cars racing around the track and found myself counting each impression specific cars made on me. Then, out of the blue, I was thinking how much it would suck to be the one who had to count those impressions. Well, they’re now thanking Nvidia and SAP.
The two tech giants are teaming up to use artificial intelligence and computer vision to figure out how many times brands appear in the real world. This is mostly still done by people, who count, or at least estimate, how many advertisement impressions are made. Nvidia showed off their new technological breakthrough in a demo at its GPU Technology Conference in San Jose, California.
It was only this year that SAP, using Nvidia’s new DGX-1 AI platform with eight Tesla V100 AI processors, announced that they would be increasing its machine learning and artificial intelligence develop efforts in order for its applications will have a broader reach when it comes to automating processes such as payment processing, sales discounting and employee approvals.
Now using computer vision, now SAP AI says it can do that job better. It’s called Brand Impact and is able to calculate the exact impact exposure has on a brand in the real world. The appearance of ads or a logo on a race car in a race broadcasted on television can be seen by computer vision and using deep learning neural networks.
“It makes so much sense for SAP to be working on deep learning,” said Jen-Hsun Huang, CEO of Nvidia. “90 percent of world’s largest enterprises are in the SAP enterprise resource planning (ERP) system. If we could figure out how to use AI to find insight in that dark matter, it would be incredible.”
At the VentureBeat gathering, MB 2017, Amazon, Google, IBM, Facebook and other leaders of AI were under one roof with other brand bosses such as The New York Times, Walmart, Coca-Cola and Tumi, as well as rising star startups like Visabot, Mezi, Bark.us and Octane AI. They were all saying the same thing: there is a huge need for a practical, yet efficient AI.
For example, Walmart employs ML to improve its service for its 140 million shoppers that visit their stores weekly, making new services possible and offering coupons and promo codes based on their data. Vice president of customer experience engineering (never knew there was such a position) at WalmartLabs said the role of data science is to make so-called Pick-up Towers within stores possible.
These Pick-up Towers allow customers to order and pay online for items and then simply come to the store and pick them up, totally skipping long checkout lines. They are also testing a 20-store pilot of Scan-and-Go shopping, which is much like Amazon’s cashierless convenience store.
Improving customers’ shopping experience was just one way in which ML is being used by Walmart. The retail giant recently opened an “associate delivery program” to defeat the “last mile” problem. In order to do this, Desegur built a system that looks much like Uber’s.
It determines who is best able to deliver to which customer, but it must also be layered in with inventory management to be certain only the desired merchandise is available for delivery with the push of a button. This will totally change the game for Walmart and make it tough to compete against, for sure. And all thanks to our friends, the learning machines.