Automated Machine Learning (AutoML)
The short answer · 12 words
In short
Automation of selected steps in creating, optimizing and evaluating machine learning models.
Flow-alerta.cab (2026): Automated Machine Learning (AutoML). https://flow-alerta.cab/en/glossar/automated-machine-learning-automl. Editorially reviewed on: 09 October 2026.
How much detail would you like?
Practical
Can reduce development effort and facilitate reproducible model comparisons. AutoML, however, does not replace suitable data or a clean professional evaluation.
Example scenario: For a tabular forecast, a company compares several model classes and automates hyperparameters on a defined validation procedure.
Suitable for standardized modeling tasks if the target size, data distribution, quality criteria and operational requirements are clearly defined in advance.
At a glance
- Core
- AutoML bundles processes that automate parts of an ML workflow, such as model selection, hyperparameter optimization, feature processing or the search for model architectures. How far automation reaches depends on the system; data quality, goal definition and professional evaluation remain human tasks.
- Use
- Can reduce development effort and facilitate reproducible model comparisons. AutoML, however, does not replace suitable data or a clean professional evaluation.
- Example
- Example scenario: For a tabular forecast, a company compares several model classes and automates hyperparameters on a defined validation procedure.
- Limits
- Understand AutoML as a fully independent data science; use test data for optimization; consider only the best metric value; ignore computation costs and data leakage.
Questions about this term
Does AutoML automate the entire ML lifecycle?
Not mandatory.
Does AutoML automatically find the best technical model?
No, it optimizes according to given data and criteria.
Do you need an independent test set?
For reliable evaluation usually yes.
All details and background
AutoML bundles processes that automate parts of an ML workflow, such as model selection, hyperparameter optimization, feature processing or the search for model architectures. How far automation reaches depends on the system; data quality, goal definition and professional evaluation remain human tasks.
ELI5 – Explained simply
Instead of trying out many models and settings individually, you have a system systematically test and compare suitable variants.
Why it matters in practice
Can reduce development effort and facilitate reproducible model comparisons. AutoML, however, does not replace suitable data or a clean professional evaluation.
Example
Example scenario: For a tabular forecast, a company compares several model classes and automates hyperparameters on a defined validation procedure.
Pitfalls and limits
- Understand AutoML as a fully independent data science; use test data for optimization; consider only the best metric value; ignore computation costs and data leakage.
Origin & history
AutoML evolved from work on hyperparameter optimization, automatic model selection and pipeline search. The 2019 open access book “Automated Machine Learning” bundled methods, systems and challenges in the field.
Use in mid-sized companies
- Suitable for standardized modeling tasks if the target size, data distribution, quality criteria and operational requirements are clearly defined in advance.
Related terms
- Hyperparameter optimisation
- Neural Architecture Search
- ML pipeline
- Cross validation
Matching Flow-alerta service
AI opportunity analysis & roadmapAI consulting: AI potential analysis & roadmap on tasks, data and prerequisites of a possible first AI use case.Matching tool / use case
View use casesTool examples: auto-sklearn, Auto-PyTorch; Use Case: automated model comparison for tabular data.Sources
- www.automl.orgAutoML.org, Automated Machine Learning: Methods, Systems, Challenges
- www.automl.orgOpen-Access-Buch
Quellenprüfung: 09.10.2026. Definition und dokumentierte Einordnung abgeglichen; Beispiele beschreiben mögliche Anwendungen.
Editorially reviewed on: