AWS, Veritone: Machine Learning, AI Can Help Monetize Ads, Content

Machine learning (ML) and artificial intelligence can be used by media and entertainment (M&E) companies to help them monetize broadcast ads and content, according to Amazon Web Services (AWS) and Veritone.

“At Amazon, we have been making investments in machine learning for many years,” Christopher Kuthan, AWS business development lead for Media Industry Solutions, said Nov. 6 during the webinar “Monetize Media Broadcast Ads & Content with Machine Learning” (ML).

“Customers’ experiences at Amazon are driven through machine learning capabilities,” including its supply chain forecasting, fulfillment and logistics, he told listeners. Its new drone initiative is “driven by deep learning and machine learning capabilities” also, he said, noting the company has “thousands of engines…that are committed to machine learning and deep learning, and it’s really a deep part of our heritage.” A wide range of Amazon customers and partners are now using ML on the AWS platform, he noted.

Amazon’s recommendation engine, meanwhile, has already been using AI “to provide the right recommendation to our customers” for many years, he said.

“Our mission is to take our rich experience and expertise with machine learning across Amazon and AWS, and put it in the hands of all organizations – so, democratize it,” he said, adding: “Every developer, data scientist and researcher should have the capabilities at their fingertips that machine learning and deep learning gives them, and build really intelligent applications based on those.”

Amazon’s ML stack provides a broad, deep set of capabilities that can help put AI to work for a company’s business, according to Kuthan. AWS can also help companies implement ML for its specific use cases, he said, adding the technology is still too complicated for everyday developers.

Amazon’s SageMaker ML platform was designed to help companies overcome the challenges of deploying ML, he noted. He went on to cite the National Football League (NFL) as an example of a customer using SageMaker. One week of NFL games now creates 3 TB of data and the NFL uses SageMaker to analyze telemetry data to predict plays, according to Amazon. As a result, computations that could take months to refine now take only weeks or days, Amazon said.

Veritone’s aiWARE AI platform, meanwhile, helps that company’s customers by transforming and analyzing their assets, including audio and video, “into actionable insights,” according to Drew Hilles, its SVP of M&E. The company now has more than 600 clients across the M&E, legal & compliance, and federal government and state, local government and education sectors, he said.

“We do know that 80 percent of the world’s information is locked without cognition,” including phone calls and other audio, as well as video – and “all unstructured” where “we don’t know what’s in the content,” he said. “We also know that 90 percent of the world’s data was created in the last 24 months,” he said, adding the open, cognitive operating system that Veritone created was designed to “help unlock that 80 percent” of data.

During the webinar, listeners were informed how they can: automate ad verification and reporting tasks; enhance archive content to make media searchable and reusable; use AI and ML in the cloud for near real-time media intelligence; and start applying ML tools.

Hilles and Charles Steinhauer, COO of U.S. radio syndication giant Westwood One, also explained how Veritone’s aiWARE, paired with AWS, helped Westwood One develop AI and ML solutions for ad verification and monetizing archived media. 
Outside the M&E space, Hilles also said the aiWARE application Identify was recently used by a Southern California sheriff’s department to quickly identify and catch a “sexual predator…within hours of the crime” being committed.