Amazon launched its Sagemaker tool for the APAC region last week, in a bid to take the complexity out of the machine learning workflow. The tool is built for companies that are either unfamiliar with ML or don't have the time to train their own models.
75% of the content on video-on-demand platforms is driven by recommendation-based machine learning, said Mai-Lan Tomsen Bukovec, VP and general manager of Amazon S3, in her keynote address at the first AWS Summit in Singapore. Bukovec also offered examples of businesses that rely on Amazon to achieve scale and speed with machine learning.
"Right here in Southeast Asia, Sunday Insurance uses machine learning algorithms to customise insurance policies for customers," she said, adding that risk is analysed at a scale which results in 10 to 20% lower premiums than traditional insurers.
She also talked about Shopper, an outfit recommendation engine that uses Amazon Recognition to analyze over 100 million social media images every month.
"When you have a significant data set, you often have a significant opportunity to try learning off that data set to help your business, to help customer experience, and it has incredible potential that is in many ways untapped," she said.