Democratizing deep learning in the cloud

November 20, 2018 - Research S177

With the cloud enabling improvements to efficiency, collaboration, and access to data, businesses can employ deep learning at scale in a way that was not possible using a traditional on-premise deployment model. Now, companies who lack the funding or resources to build and maintain on-premise infrastructure for deep learning projects are able to leverage the technology via the cloud. Nucleus interviewed more than 30 individuals carrying out 140 distinct deep learning projects to examine how prevalent the cloud is in modern deep learning execution and the benefits users received from deploying in the cloud rather than on-premise. Nucleus found that companies can reduce the initial cost of their deep learning project by 45 percent and increase overall efficiency by 20 percent by using the cloud.