Google's TensorFlow Powers Data Robot's Machine Learning Platform

by Ostatic Staff - Nov. 02, 2016

Artificial intelligence and machine learning are creating a lot of buzz right now, and open source tools are part of the buzz. A few months back, Google made a hugely influential contribution to the field of machine learning. It open sourced a program called TensorFlow that is now freely available. It’s based on the same internal toolset that Google spent years developing to support its AI software and other predictive and analytics programs.

Now, data science company DataRobot has announced the latest version of its enterprise machine learning platform. The new release integrates the TensorFlow library for deep learning along with new tools to help users extract insights from all models on the platform. This is a great example of how powerful open source tools are driving commercial offerings forward.

"For years, the lack of visibility into why a prediction decision is made has been a big issue for making machine learning accessible to anyone but data scientists with PhDs," said Phillip Gurbacki, DataRobot VP of Product Management. "Now with DataRobot Reason Codes, users across a business - not just data scientists - have answers for every prediction.  With every update, we're continually raising the bar on what it means to be an enterprise machine learning platform, and delivering a decided competitive advantage for our customers."

The latest version of the DataRobot platform offers new features, including:

Deep Learning with TensorFlow:  TensorFlow is the open source library Google develops and maintains to support deep learning as the "brain" in many of their online services. By packaging the power of TensorFlow in an easy-to-use and automated platform, DataRobot allows anyone to optimize their business with deep learning.

Reason Codes:  This innovative new feature allows users to see the exact evidence a model relied on for any prediction.  Reason codes provide insights to the users building models, and clear explanations to anyone making decisions based on the model's predictions.

Feature Impact: Feature impact allows users to create simple graphs showing the primary drivers of their key business outcomes. Feature Impact is useful for everything from iterative model improvements to strategic decision-making by senior executives.

DataRobot Prime: A premium add-on, DataRobot Prime allows users to create Python or Java code approximating any model with a single click, so models can be deployed in almost any environment.

"Visionary companies seek ways to break machine learning out of the 'expert' silo and deploy it across the enterprise," said Thomas Dinsmore, author of Disruptive Analytics and independent industry analyst. "DataRobot uniquely offers the simplicity, speed, scalability, and accuracy needed for enterprise machine learning. This latest version, with TensorFlow support and improved diagnostics, enables a broad pool of users to take advantage of advanced predictive analytics and deliver better predictions faster."

Interested in more on machine learning? We've done extensive coverage of some of the recently open sourced tools in this arena here