dotData Updates dotData Enterprise and dotDataPy Data Science Acceleration 

dotData, a provider of data science automation and operationalization, has released Version 1.6 of dotData Data Enterprise and Version 1.2 of dotDataPy. The new updates provide users with deeper insights, increased flexibility, and greater security.

"One of the most exciting enhancements in the update is the ability to automatically generate features from text data in combination with other types of data sources. This new feature unlocks the tremendous value of in-house business text data owned by many enterprises," said Ryohei Fujimaki, founder and CEO of dotData, in a statement. "Another notable enhancement is support of deep learning frameworks such as TensorFlow and PyTorch in our AutoML with enhanced transparency for highly non-linear ML models."

Key updates of the dotData Enterprise Version 1.6 and dotDataPy Version 1.2 include the following:

• AI-powered feature engineering for text data--dotData now supports automated feature engineering for text data and unlocks text information, such as call center customer feedback, sales reports, meeting minutes, geo-locational text data, and temporal text data as well as traditional static text data.
• Deep learning in AutoML--dotData now supports  deep learning frameworks such as TensorFlow and PyTorch, as part of its AutoML capability so users can test advanced neural network methods. Additionally, advanced feature validation methods, such as permutation importance, are now supported and available both on dotData GUI and dotDataPy.
• More flexible and secure Hadoop deployment--dotData Enterprise Version 1.6 now supports more flexible deployment on existing Hadoop clusters. Additionally, in addition to Kerberos (supported in dotData Enterprise 1.4), dotData 1.6 supports more advanced Hadoop security options to protect enterprise data.
• Progress visualization for greater transparency of task status.
• The Model Insights dashboard for comparisons of many machine learning models to analyze model behavior deeply.