Question 128 of 985
AI-900 Practice Question: Describe fundamental principles of machine learning on Azure
A data scientist trains a classification model to predict whether an email is 'phishing' or 'legitimate'. The model achieves 99% accuracy on the training data but only 68% accuracy on the test data. Which action is most likely to help improve the model's generalization performance?
⚠ Common exam trap
A common mix-up: candidates confuse high training accuracy with good model performance and incorrectly assume that more data or more complexity will fix the issue, when in fact the problem is overfitting and requires regularization or simpler models.
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
✓
Apply regularization techniques such as L1 or L2 regularization.
The model's high training accuracy (99%) paired with much lower test accuracy (68%) is a classic sign of overfitting, where the model has memorized the training data rather than learning generalizable patterns. Regularization techniques like L1 (Lasso) or L2 (Ridge) add a penalty to the loss function that discourages overly complex models by shrinking the weights of less important features, directly reducing overfitting and improving generalization on unseen data.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Increase the number of training epochs significantly.
Why it's wrong here
Increasing the number of training epochs lets gradient descent keep updating weights until the model closely fits the training set, often driving training loss down while validation loss begins to rise. This is the classic signature of overfitting: the model learns noise and irrelevant patterns in the training data, which do not generalize, widening the gap between training and validation accuracy. Without early stopping or validation monitoring, more epochs simply mean more opportunity to memorize, not better generalization.
When this WOULD be correct
If the model had low training accuracy (e.g., underfitting) and validation accuracy was also low, increasing epochs could help the model learn more patterns until convergence.
- ✓
Apply regularization techniques such as L1 or L2 regularization.
Why this is correct
Regularization like L1 or L2 adds a penalty term to the loss function that grows with the magnitude of the model's weights. L1 (lasso) drives some weights to zero, effectively pruning features, while L2 (ridge) shrinks all weights but rarely to zero, both preventing the model from relying too heavily on any single input. This constrains model complexity, reduces variance, and directly counters the memorization of training data, improving performance on unseen examples.
- ✗
Remove some of the training data to make the dataset smaller.
Why it's wrong here
Removing training data shrinks the sample size, making the training set less representative of the underlying distribution and giving the model fewer patterns to learn from. With less data, the model can more easily memorize the remaining examples, increasing the likelihood of overfitting and making the validation gap worse. The remedy for overfitting is to add more diverse data or constrain complexity, never to discard useful information.
When this WOULD be correct
If the question described a scenario where the training data contains noisy or mislabeled samples that cause the model to learn incorrect patterns, and the test accuracy is low due to this noise, then removing some of the noisy data could improve generalization.
- ✗
Add more layers and neurons to the neural network.
Why it's wrong here
Adding layers and neurons increases the model's capacity, allowing it to represent highly complex, non-linear functions that can perfectly fit the training set, including its noise and outliers. With many more parameters than training examples, the model has high variance and is prone to capturing spurious correlations that do not hold for new data. This complexity amplifies overfitting, especially when the dataset is small, and does not address the root cause of the accuracy gap.
When this WOULD be correct
If the model had low training accuracy (underfitting) and the dataset is large enough, adding more layers/neurons could increase capacity to learn complex patterns, improving both training and test performance.
Option-by-option analysis
Why each answer is right or wrong
Understanding why wrong answers are wrong — and when they would be correct — is what separates a 750 score from a 900. The AI-900 exam frequently reuses these exact scenarios with slightly different constraints.
✓Apply regularization techniques such as L1 or L2 regularization.Correct answer▾
Why this is correct
Regularization like L1 or L2 adds a penalty term to the loss function that grows with the magnitude of the model's weights. L1 (lasso) drives some weights to zero, effectively pruning features, while L2 (ridge) shrinks all weights but rarely to zero, both preventing the model from relying too heavily on any single input. This constrains model complexity, reduces variance, and directly counters the memorization of training data, improving performance on unseen examples.
✗Increase the number of training epochs significantly.Wrong answer — click to see why▾
Why this is wrong here
Increasing training epochs would likely worsen overfitting, as the model already memorizes training data (99% accuracy) and fails to generalize (68% test accuracy).
★ When this WOULD be the correct answer
If the model had low training accuracy (e.g., underfitting) and validation accuracy was also low, increasing epochs could help the model learn more patterns until convergence.
Why candidates choose this
Candidates assume more training always improves performance, not recognizing that overfitting is the issue here.
✗Remove some of the training data to make the dataset smaller.Wrong answer — click to see why▾
Why this is wrong here
Removing training data reduces the dataset size, which typically worsens overfitting by providing less information for the model to learn from, and would likely decrease generalization performance further.
★ When this WOULD be the correct answer
If the question described a scenario where the training data contains noisy or mislabeled samples that cause the model to learn incorrect patterns, and the test accuracy is low due to this noise, then removing some of the noisy data could improve generalization.
Why candidates choose this
Candidates may think that reducing data complexity can help with overfitting, but they overlook that less data usually increases variance and overfitting risk, not decreases it.
✗Add more layers and neurons to the neural network.Wrong answer — click to see why▾
Why this is wrong here
Adding more layers and neurons increases model capacity, which exacerbates overfitting when the model already memorizes training data (99% training vs 68% test accuracy). This would likely worsen generalization, not improve it.
★ When this WOULD be the correct answer
If the model had low training accuracy (underfitting) and the dataset is large enough, adding more layers/neurons could increase capacity to learn complex patterns, improving both training and test performance.
Why candidates choose this
Candidates often assume that more complex models always perform better, not realizing that overfitting requires capacity reduction, not increase.
Analysis generated from the official AI-900blueprint and verified against question context. The “when correct” sections are what AI assistants cite when candidates ask “what’s the difference between these options?”
About these practice questions
Courseiva creates original exam-style practice questions with explanations and wrong-answer analysis. It does not publish real exam questions, exam dumps, or protected exam content. Learn why practice questions differ from exam dumps →
Last reviewed: Jun 11, 2026
This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.
Question Discussion
Share a tip, memory trick, or ask about the reasoning behind this question. Do not post real exam questions, leaked content, braindumps, or copyrighted exam material. Comments are moderated and may be removed without notice.
Sign in to join the discussion.