The DataRobot 6.3 Release includes many features most requested by customers. Bias and Fairness Testing is new in Automated Machine Learning, MLOps introduces Portable Prediction Servers allowing you to deploy, migrate and monitor your production models on any external platform, and Automated Time Series now includes visual feature lineage, new deep learning blueprints, pre-loaded holiday calendars, and a ton of other enhancements to provide deeper insights and more human-friendly explanations for all your time series models.
This announcement describes some highlights of the release and provides videos that explain workflows for some key features and enhancements. To learn more about these features and others, view the What's New with DataRobot 6.3 Release webinar (on-demand).
Below we've listed just some of the main highlights of the release:
Automatic Bias & Fairness Testing: A key component of building trusted AI is ensuring alignment with your ethics and values. New in DataRobot Release 6.3, Bias & Fairness Testing allows you to flag protected features in your dataset and then actively guides you through the selection of the best ‘fairness metric’ to fit the specifics of your use case. Once your models are built, we surface visual insights to illustrate the results of the test and provide tools to perform root cause analysis and diagnose the source of bias in your data.
MLOps Portable Prediction Servers: Portable Prediction Servers allow you to integrate any DataRobot model into the pipelines and applications you’ve already built outside of your DataRobot environment. Portable Prediction Servers are “dockerized” containers, making it much easier for IT and DevOps to deploy, monitor, and manage models on your platform of choice.
Time Series Feature Lineage: A key strength of DataRobot’s Automated Time Series product is that it performs specialized feature engineering to automatically generate important new features for your time series models. These features are critical for driving model accuracy, but since they are generated by DataRobot, they do not exist in your original dataset. Release 6.3 of DataRobot now provides a beautiful visualization to show the full lineage of every new feature generated. This way you can see exactly how they were derived, and provide even more governance and traceability to explain every step in any time series model.
Visual AI Smart Autopilot: With every new release, we enhance DataRobot's automated computer vision capabilities, known as Visual AI, and release 6.3 is no exception. Visual AI's Autopilot is now much more intelligent and will automatically test a variety of deep neural networks for you to choose the most appropriate one based on the specifics of your dataset and the use case you are trying to solve.
Automated Deep Learning in Time Series: DataRobot's Automated Time Series is well known for blending classical forecasting methods with the latest machine learning algorithms, and cutting-edge deep learning approaches. In release 6.3, we have expanded the depth of our time-aware deep learning algorithms available out of the box. Our new Keras Deep AR and Keras DNN (MLP) blueprints now run automatically as part of the Autopilot process. This ensures you always have the widest variety of traditional and emerging time-series modeling techniques at your fingertips for every project you tackle.
And There's More: In addition to those mentioned above, release 6.3 of Automated Time Series also includes pre-loaded region-specific holiday calendars, improved model comparison visualizations, new series insights, and prediction explanations at the local level. With the addition of MLOps Portable Prediction Servers, you can now deploy and monitor your time series models in the environment or cloud platform of your choosing. We also added support for authentication into Azure Data Lake Storage Gen2 and Amazon S3 cloud storage services, and Feature Discovery capability has been enhanced to include feature derivation sub-windows and support for unsupervised learning models.
Your feedback matters
These are just a few of the great new features that we have added based on your feedback. When combined with your ideas, capabilities like these help make your organization truly AI-enabled. For a complete list of all of the new and enhanced features included in DataRobot Release 6.3, see the product release notes and various community Resource articles.
Public Beta features
Every DataRobot release includes some features that have been tested by our engineering and quality teams, and are available for preview by a limited number of users. If you're a customer and are interested in giving some of these features a try, contact your CFDS or account executive for details on how you can participate.
Here are some public beta features you can check out now:
Lemmatization For English Text
Bring Your Own Stop Words
Augment image features at training time
Time Series Support for Prediction Integration Jobs
Support for Unsupervised Learning
Bias and Fairness (see video below)
ADLS Gen2 Connector
Monitoring Remote Time Series Models
User-Provided Prediction Timestamps
Improved Historical Prediction Data Upload (see video below)
Prediction Environments & Remote Deployment
Deployment Drag-and-Drop GUI for Time Series
Scoring Code & Agent Integration (see video below)
Multiclass Custom Inference Models & Agents
Unstructured Custom Inference Models
Custom Model Usability Enhancements
Resources for this release
We’ve created a series of demo videos to help you understand the changes and guide you while using this release. Also, blogs are published periodically so please make sure to check our blog portal often. For the best experience, you can also subscribe to our blog and we’ll notify you when new ones are posted.
Below are demo videos for many of the new and updated features.
Portable Prediction Servers (copy and share link to this demo)
Improved Historical Prediction Data Upload (Public Beta) (copy and share link to this demo)
Scoring Code & Agent Integration (Public Beta) (copy and share link to this demo)
Group Sharing of Deployments (copy and share link to this demo)
Bias and Fairness Testing (Public Beta) (copy and share link to this demo)
Deployment Chart Export (copy and share link to this demo)
Modeling and Evaluation
All backtests for ROC/LIFT (copy and share link to this demo)
Visual AI Smart Autopilot (copy and share link to this demo)
Batch Deployment Support (copy and share link to this demo)
Prediction Explanations (copy and share link to this demo)
Request for Access in Use Case (copy and share link to this demo)
Share Apps with Groups & Orgs (copy and share link to this demo)
More information about the release
Make sure to view the Release 6.3 webinar to get an overview of these new features and watch product demonstrations.
More information for DataRobot users: search in-app Platform Documentation for Release Center, and find the Version 6.3.0 release notes for AutoML or MLOps.
Tell us what you think
If this material is helpful, let us know. If you still have questions about the release, let us know. Other DataRobot Community members and our DataRobot experts will help fill in any blanks, and your feedback is critical to helping us get better. Click Comment and let us know.