Describe fundamental principles of machine learning on Azure →mediumMultiple ChoiceObjective-mapped

AI-900 Practice Question: Describe fundamental principles of machine learning on Azure

What is 'regularization' in machine learning and why is it used?

⚠ Common exam trap

Candidates often confuse regularization with data normalization or standardization, because both involve 'regularizing' data in a colloquial sense, but regularization is a penalty on model complexity, not a data transformation step.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Adding a complexity penalty to the training objective to reduce overfitting

Regularization is a technique used to reduce overfitting by adding a penalty term to the loss function during training. This penalty discourages the model from learning overly complex patterns (e.g., large weights) that fit the training data too closely but fail to generalize to new data. In Azure Machine Learning, regularization can be applied via algorithms like Lasso (L1) or Ridge (L2) regression, which directly modify the optimization objective.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
Normalizing input data to a standard scale before training
Why it's wrong here
Normalizing input data to a standard scale, such as z-score or min-max scaling, is a preprocessing step meant to place features on comparable ranges so gradient-based optimization converges more reliably. It changes the representation of feature values, but it does not introduce any term into the loss function that penalizes large parameters. Scaling improves numerical stability and can indirectly affect weight magnitudes, but it is not regularization because it never constrains the model's complexity in the training objective.
✓
Adding a complexity penalty to the training objective to reduce overfitting
Why this is correct
Adding a complexity penalty to the training objective reduces overfitting by discouraging overly large parameter values through an explicit term in the loss function. In L2 regularization, the squared magnitude of weights is added to the loss, while L1 regularization adds the absolute values of weights, promoting sparsity. This penalty biases the model toward simpler, more generalizable hypotheses, directly addressing the variance component of the bias-variance tradeoff. Therefore, this is the accurate description of regularization.
✗
Ensuring models comply with AI regulations in different jurisdictions
Why it's wrong here
Regularization in machine learning refers to mathematical techniques applied directly to the training objective, such as adding an L1 or L2 penalty on model weights to discourage complexity. Ensuring models comply with AI regulations in different jurisdictions is a legal and governance activity involving data protection laws, ethical frameworks, and auditing requirements. Regulatory compliance does not alter the loss function or the optimization process during training, so it is not an example of regularization.
✗
Standardizing the format of training data from different sources
Why it's wrong here
Standardizing the format of training data from different sources is a data engineering task that ensures schemas align, columns have consistent units, and missing values are handled uniformly before training begins. Regularization, by contrast, operates within the loss function during model training and involves no modification to the raw data or its schema. Although consistent data formats facilitate model training, they do not add a complexity penalty, so they are unrelated to regularization.

Go deeper

AI-900 Practice Question: Describe fundamental principles of machine learning on Azure

Why each option matters

Option-by-option breakdown

Related to this question