Release notes AutoML v.0.3.0: user management and new data science modules

We’re excited to ship the latest release of Modulos AutoML: you can now add and manage users and make use of a wider range of data science modules.

User management

We added the user management functionality so that your whole team can use the same AutoML instance and keep track of who uploaded which data set and who started which workflow.

AutoML now has two classes of users:

  • Administrators, who can see and modify anything, and
  • Users, who can see and modify only data sets and workflows that they own

Data science modules

AutoML was conceived to be completely modular so that we could easily adapt a wide range of data science modules.

In this release, we’re adding quite a few:

  • Two new models to cover the area of regression tasks: 
    • A deep learning regressor model
    • the XGBoost regressor model
  • Several new feature engineering modules:
    • A PCA feature extractor,
    • An Autoencoder image feature extractor
    • A T-test feature extractor

Together, these new data science modules extend the range of tasks which you can tackle with AutoML and reach top performance.

Other improvements and bug fixes

  • You can now see a progress bar during data set validation. The validation of large data sets can take a while, and previously there was no indication of progress.
  • We improved our GPU processing power while still keeping it as an experimental feature, trying to get the most out of it of you.”
  • The on-platform documentation has been significantly expanded.