NCA-GENL Core Machine Learning and AI Knowledge Practice Question
A machine learning engineer is training a convolutional neural network for image classification and notices that the training loss decreases steadily, but the validation loss starts increasing after a few epochs. The training set is large and representative. Which technique is most directly aimed at addressing this phenomenon?
⚠ Common exam trap
Test-takers frequently confuse overfitting with optimization issues, leading to learning rate or batch size tweaks instead of proper regularization.
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
✓
Add dropout layers
The described behavior—training loss falling while validation loss rises—is classic overfitting. Dropout is a direct regularization technique that prevents co-adaptation of neurons and improves generalization. Increasing learning rate, reducing training data, or changing batch size do not specifically counter overfitting and may worsen the outcome.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Reduce the size of the training set
Why it's wrong here
Shrinking the training set would reduce the amount of data available to learn from, likely increasing both training and validation loss and harming generalization. The problem is not insufficient data but the model's excessive capacity relative to the data's complexity. Removing data does not address the overfitting mechanism and is generally poor practice.
- ✓
Add dropout layers
Why this is correct
Dropout randomly deactivates neurons during training, forcing the network to learn redundant representations and reducing its ability to memorize training noise. This regularization directly combats overfitting, which is the cause of rising validation loss despite decreasing training loss. In this image classification scenario, dropout is a standard and effective remedy when the model has sufficient capacity to overfit.
- ✗
Use a smaller batch size
Why it's wrong here
Batch size affects the noise in gradient estimates and can influence convergence speed and generalization slightly, but it is not a direct regularization method for overfitting. A smaller batch size may add stochasticity that acts as a mild regularizer, yet it is far less targeted than explicit techniques like dropout or weight decay. In this scenario, it would not reliably stop validation loss from rising.
- ✗
Increase the learning rate
Why it's wrong here
Increasing the learning rate would likely make the divergence worse by causing larger parameter updates that overshoot minima. While a too-small learning rate can slow convergence, it does not cause validation loss to rise while training loss falls. The observed pattern is overfitting, so adjusting the learning rate upward is counterproductive here.
About these practice questions
One of 367 original NCA-GENL practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official NVIDIA exam blueprint
This NCA-GENL practice question is part of Courseiva's free NVIDIA 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 NCA-GENL exam.