If computing is to realize its full potential, scientists will need new lenses for the digital eye. It’s as though human vision’s technologies, seen through an ever-shortening window, need to be reshaped. Computers may increasingly surpass human perceptual abilities (see What Computers Really Can Do Below), but top executive, teams and vision experts are now mapping the future of vision and marketing along with AI in a new book I authored, Resurrecting Vision, Globalizing Amazon, the World’s Largest Retail Company.
Alex Tse, the Chief Scientist of Amazon, knows well the path that once followed. From 2014 to 2017, Tse led a team of engineers who gathered user feedback about Alexa in the Amazon Echo — a game-changing innovation in voice personalization that transformed the way users interact with their devices, like the remote control for the bedroom TV.
These days Tse concentrates his time in machine learning and deep learning, which “are delivering great innovations in big-picture research on all our products,” he says.
Tse’s work revolves around linking data on the content and speed of a product’s production to the search data with which you can interact. Machine learning plays a key role here, applying the complexity of a wide variety of human experiences (such as human focus, attention, predisposition) to a narrow set of signals, in order to understand how the features of Amazon products, such as Amazon delivery, versus a rival brand’s offerings are likely to affect a consumer’s desire to purchase. It helps to understand the user, in other words.
This kind of consumer psychology has bedeviled computer science for a long time, but we are at a turning point where innovative uses of vision and imaging can deliver new levels of logic, precision and optimal decision making, helping brands take ownership of the customer experience.
Here’s a quick look at what the future of computer vision holds.
Internet of Things
“At AWS we see it happening from the front end of building cloud services to running cloud services,” says Tse. “We are looking at how we can make software work better.” This translates to this year, when he helped build AWS Cloud Vision, the world’s first fully distributed visual computing platform. Today, AWS companies such as Adobe, Microsoft, Bitcasa and SendGrid use Amazon Cloud Vision’s scalable services to develop visual applications, developing enterprise applications of computer vision at scale.
Intelligent Vision enables engineering and manufacturing, as well as machine learning labs in the not-too-distant future.
In 2015, Tse helped create Zero-Light Energy (ZL) lab in Redwood City, CA where AWS platforms for machine learning and artificial intelligence are tested for natural light data, so engineers can measure indoor lighting in facilities such as Fab Labs, adding a dimension to conventional light meter measurements of outdoor lighting that can inform local LED lighting options.
Of course, companies are already more proactive, more intently focused on their customers.
I just saw the graphics for Amazon Robotics. When I looked at these screens, the entire field is being automated, and even dogs like Princess and Fred are going to be doing it. pic.twitter.com/UGjUz3hZR1 — Bernard Marr () June 15, 2018
For instance, British airline Monarch restructured its fleet and workforce to become leaner and more nimble. The airline is even embracing to avoid job losses, including hand-picking new employees when flying. In an interesting stunt last week, it announced that it was getting rid of mechanical check-in counters, instead moving forward with automated luggage verification kiosks.
Every employer we meet is moving towards flexibility, according to Tse. That means a company like Facebook, whose recently opened Burj Khalifa cultural center, features a 1,600-person open-air amphitheater capable of taking the pressure off staff:
This works best for a big event, but you can also put a temporaryized event together like a mini Cannes Lions. Everyone we know does events in the middle of summer, and one of the big issues we hear about from attendees is what to do with all the spare baggage they have. If a place like Facebook can use technology like this to handle that, it’s a win for everyone.
AI and Machine Learning