GraphLab Becomes Dato, Raises $18.5M for Machine Learning Tech
GraphLab, a provider of analytics software based on open source code, is changing its name to Dato as it aims to tackle the growing demand for predictive applications that can leverage machine learning.
The Dato name change is being accompanied by new funding. The company just announced it has raised Series B funding of $18.5 million, bringing the total capital raised to $25.25 million.
The open-source GraphLab graph analytics project is at the core of Dato. The company, which is led by CEO and co-founder Carlos Guestrin, an Amazon Professor of Machine Learning in Computer Science and Engineering at the University of Washington, also offers a commercial product called GraphLab Create that was first announced in October.
“While we began life as a graph analytics open source project, the underlying engine has morphed a couple of times, adding significant innovations for the handling and analysis of tabular data as well as text and images,” Guestrin told Enterprise Apps Today.
Dato calls the tabular data structures SFrames, explained Guestrin, noting that they provide a significant performance advantage over alternatives in computational scale. For advanced machine learning, Dato offers a deep learning module which enables highly accurate recognition of images.
“The list really goes on, and our new name is apt to fit both our current and planned solution enhancements,” Guestrin said. “Dato literally means ‘unique data item' in Portuguese and Spanish, my native languages.”
Intelligent Apps
While GraphLab has a new name and new funding, Guestrin said the direction is the same as it has been since the company was incorporated in March 2013 – to deliver the world's best platform for creating intelligent (i.e. predictive) applications.
The flagship product has been adding users at a “super brisk clip,” Guestrin said, and has been downloaded thousands of times since general availability and pricing was announced in October.
There is a direct intersection between Dato's tools and the growing market for Big Data solutions, and in particular Hadoop. Guestrin said Dato supports all data types, including tables, graphs, images or text. It can access data for analysis whether it resides on a laptop, or in a file system, a database, Hadoop distribution or a private/public cloud environment.
“Many of our customers have made an investment in Hadoop for their data aggregation needs and so we support the various popular distributions,” Guestrin said, noting that Dato's platform is Cloudera CDH5 certified.
Machine Learning Analytics Challenges
Machine learning is the discipline by which applications are made intelligent, Guestrin said. However, a number of challenges must be solved before machine learning gains widespread acceptance.
“The biggest challenge is leveraging machine learning to create those predictive apps and making them broadly available to customers and users,” Guestrin said.
In many organizations, the people tasked with building predictive applications either don't have the data science skills required to build the apps or lack the knowhow and resources to scale them into responsive services.
That's the challenge Guestrin aims to overcome with Dato. “A single data scientist or programmer or app developer with a laptop can build a predictive application that drives revenue, and do it fast,” he said.
Sean Michael Kerner is a senior editor at Enterprise Apps Today and InternetNews.com. Follow him on Twitter @TechJournalist.
Sean Michael is a writer who focuses on innovation and how science and technology intersect with industry, technology Wordpress, VMware Salesforce, And Application tech. TechCrunch Europas shortlisted her for the best tech journalist award. She enjoys finding stories that open people's eyes. She graduated from the University of California.
