For several years now, many IT organizations have been building up massive repositories of Big Data based on Hadoop. Datameer today announced it is making it simpler to apply deep learning algorithms in the form of open source Google TensorFlow software within an analytics application that runs on top of Hadoop.
John Morrell, senior director of product marketing at Datameer, says a new SmartAI platform embeds a deep learning engine within a data preparation and analytics platform in a way that promises to democratize access to AI technologies.
“This is how people across the organization are going to operationalize AI,” says Morrell.
Rather than being dependent on data scientists to take advantage of deep learning algorithms, Morrell says TensorFlow has been embedded into the runtime engine of the Datameer platform. That platform was originally developed to provide an analytics application for business analysts and other end users on top of Hadoop that did not require the skills of a data scientist to invoke. SmartAI extends those capabilities now to deep learning algorithms in the form of Google TensorFlow.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=i
The primary reason there has been such a rapid advance in AI is that the cost of storing all the data needed to make advanced algorithms useful has dropped by several orders of magnitude thanks to the rise of platforms such as Hadoop. In fact, the algorithms being used to deliver those AI capabilities have been around for years. Morrell says the next step is to democratize access to those algorithms in a way that levels the competitive playing field. That can only be accomplished by embedding support for advanced AI technologies within analytics applications, says Morrell.
Precisely how those algorithms will get employed will vary considerably across organization. But as machine and deep learning algorithms become more widely accessible via applications, the less chance there is that large organizations that can afford to hire hundreds of data scientists will be able to overwhelm smaller organizations that have access to fewer, albeit sometimes savvier, business analysts and executive leadership.