Databricks ML Associate Practice Question 4: AutoML Generated Notebooks
Databricks Certified Machine Learning Associate · Test 1 · question 4 of 10
Your team is starting a new classification project. Your manager wants a strong baseline model quickly, but the data scientists insist on reviewing and changing whatever code produces the final model before it is used.
Which characteristic of Databricks AutoML addresses both requirements?
Show answer and explanation
Answer:
AutoML is a "glass box" tool. It preprocesses the data, trains and tunes several candidate algorithms, and logs every trial to an MLflow experiment. Each trial comes with a notebook of its source code. For classification and regression, the notebooks for data exploration and for the best trial are imported into the workspace, and the other trials' notebooks are saved as MLflow artifacts. The team gets a quick baseline and code it can read, clone, and change.
Option A is wrong on both counts. AutoML runs many trials across algorithms such as LightGBM, XGBoost, and random forests, and it exposes their code rather than returning an opaque model.
Option B describes a hyperparameter-tuning library such as Optuna, which tunes an objective function you write yourself. AutoML writes the training code for you.
Option D is not part of an AutoML run. AutoML produces experiments and notebooks; registering and serving a model are separate steps you choose to take.
Reference: What is AutoML?