Artificial intelligence research is still in its infancy, at least as compared to computer science in general, but the concept of unlimited computing resources is accelerating the field.
As someone with nearly unlimited computing resources at his disposal, this is something Swami Sivasubramanian, vice president of AI at Amazon Web Services, is watching play out. Last week Sivasubramanian walked GeekWire Cloud Tech Summit attendees through the array of artificial intelligence and machine-learning services that his team has developed for AWS customers and Amazon’s own internal services as well.
If you’ve been through a few tech cycles, you’ve already heard a lot about artificial intelligence. Much has been promised from this research field over several decades, but the enormous amount of data now moving into cloud computing services like AWS and others allows researchers like Sivasubramanian to make real breakthroughs that weren’t possible when data sets were scattered and siloed.
“Many of these algorithms, especially like deep learning neural nets, papers were written about even two decades ago. But what has accelerated adoption of it is that we have specialized compute infrastructure, such as GPUs, specialized CPUs, FPGAs (field programmable gate arrays), you name it,” he said. The combination of huge data sets and powerful computing engines is making AI concepts previously confined to science fiction a reality.
Take two AWS customers: CSPAN and the sheriff’s office of Washington County in Oregon. Using the company’s Rekognition image-recognition service, both were able to automate tasks that required painstaking human labor. CSPAN can now automatically identify Congresspeople speaking on the floor of the House or Senate, saving someone from having to manually annotate those videos, and Washington County is using the service to help it process photo tips when it is looking for a person of interest in an investigation.
But it’s still very early days for AI applications: Sivasubramanian joked that this world is about where the field of databases was when btrees were invented in the early 1970s. That’s about to change, however, as we gain a greater understanding of how AI models work and develop more sophisticated ways of training these systems to accomplish real goals.
Watch the full video of Sivasubramanian’s GeekWire Cloud Tech Summit talk above, and stay tuned for more highlights from the event in the days ahead.