Announcing low code Automated ML (AutoML) in Fabric Data Science
In today’s data-driven world, quickly developing and deploying machine learning (ML) models is essential. Automated Machine Learning (AutoML) streamlines this process, enabling analysts, data scientists, and domain experts to leverage ML without extensive coding. Now, with the launch of low code AutoML in Fabric Data Science, ML is more accessible than ever, regardless of technical background.
This new low code AutoML experience, now in preview, builds on the code-first version announced last April. While the code-first approach provides hands-on control, the low code version prioritizes ease of use and rapid deployment. Let’s explore the key features that make this low code AutoML experience a valuable addition to Fabric Data Science.
What’s new?
The low code AutoML experience in Fabric Data Science delivers an intuitive, guided interface for setting up and managing ML experiments, with minimal coding required. This release marks a significant step forward, providing users with the flexibility to prototype, tune, and refine models in a fraction of the time typically required. Designed with flexibility in mind, the low code AutoML interface allows users to set up end-to-end ML workflows with just a few clicks—ideal for rapid experimentation or real-world deployment.
Key Capabilities
Easily set up and configure your AutoML trial
With the low code experience, users can select their ML task—whether it’s classification, regression, or forecasting—and configure their trial according to their objectives. Users can choose from various AutoML modes that best suit their needs. For example, selecting the ‘Quick Prototype’ mode optimizes settings for a faster, more efficient training process, ideal for testing and initial evaluation. Additionally, the AutoML feature includes an auto-featurization step, which automatically generates new features for the training process, helping to improve the accuracy and effectiveness of the model with minimal user intervention.
Transparency with AutoML notebooks
One of the core benefits of AutoML in Fabric is the visibility it provides into the underlying ML processes. Each AutoML trial in the low code experience generates a fully detailed notebook that captures the trial’s configurations, workflows, and selected settings. This level of transparency enables users to walk through the steps taken during the AutoML process, making it easier to understand and explain model outcomes. Users can customize the notebook, adjust settings, and even use it to reproduce results at a later time, providing both flexibility and insight into how the model arrived at its predictions.
With ML Experiments in Fabric, all AutoML trial results are logged and stored, ensuring seamless tracking and reproducibility. The integration with MLflow allows each trial’s metrics, parameters, and model files to be saved and tracked in an organized manner. This visibility helps teams collaborate more effectively, allowing them to access run history, share models, and review key metrics at any point in the project lifecycle.
Get Started with low code AutoML in Fabric Data Science
The launch of low code AutoML in public preview marks a new chapter in democratizing machine learning. This experience brings the power of ML to users across the board—whether you’re an experienced data scientist or an analyst new to machine learning. We’re excited to provide a tool that enables users to experiment with models quickly, while maintaining transparency and control over their projects.
To try out low code AutoML in Fabric Data Science, simply access the preview through your Fabric workspace and begin your first trial today. Learn more about AutoML: Use AutoML (interface).