Question 628 of 754
Model Optimization Techniques for Faster Inference
A team is deploying a deep learning model that uses a convolutional neural network (CNN) for image recognition. The model achieves high accuracy but is very slow to infer on edge devices. Which THREE optimization techniques should the team consider to speed up inference without significant accuracy loss? (Select three.)
Quick Answer
The answer is model quantization, weight pruning, and knowledge distillation. These three techniques directly address the bottleneck of model optimization for inference speed on edge devices by reducing computational load while preserving accuracy. Quantization lowers the precision of weights from 32-bit floats to 8-bit integers, dramatically cutting memory bandwidth and accelerating matrix operations on specialized hardware. Weight pruning removes redundant connections, shrinking the model size and reducing the number of floating-point operations per inference. Knowledge distillation trains a smaller "student" network to mimic a larger "teacher" model, compressing complex decision boundaries into a lightweight architecture. On the CompTIA AI+ AI0-001 exam, this question tests your understanding of deployment trade-offs: the common trap is confusing hyperparameter tuning (which affects training speed) with inference optimization. Remember the mnemonic "Q-P-K" for Quantization, Pruning, Knowledge distillation—three levers to pull when your model needs to run fast on a tiny chip.
⚠ Common exam trap
CompTIA often tests the misconception that increasing model capacity (larger filters or more layers) improves performance without considering the trade-off in inference speed, leading candidates to select options that actually worsen latency on edge devices.
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
✓
Use weight pruning to remove unnecessary connections in the network.
Weight pruning removes redundant or less important connections (weights) from the neural network, reducing the number of computations required during inference. This directly speeds up inference on edge devices while typically causing only a minor drop in accuracy if done carefully, making it a standard optimization technique for deploying CNNs on resource-constrained hardware.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Use larger convolutional filters (e.g., 7x7 instead of 3x3) to capture more context.
Why it's wrong here
Larger filters increase computation and slow down inference.
- ✓
Use weight pruning to remove unnecessary connections in the network.
Why this is correct
Pruning reduces computation and memory footprint.
- ✓
Implement knowledge distillation by training a smaller model to mimic the larger one.
Why this is correct
Knowledge distillation creates a compact model that retains much of the original accuracy.
- ✗
Increase the number of convolutional layers to improve feature extraction.
Why it's wrong here
More layers increase computational cost and latency.
- ✓
Apply model quantization to reduce weight precision.
Why this is correct
Quantization reduces model size and speeds up inference, often with minimal accuracy loss.
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 →
Same concept, more angles
1 more way this is tested on AI0-001
These questions test the same concept from different angles. Work through them to make sure you can recognise it however the exam phrases it.
Variation 1. A self-driving car company is developing an object detection system using a convolutional neural network (CNN). The system needs to detect pedestrians and vehicles in real-time with high accuracy. Which technique can reduce inference time while maintaining accuracy?
hard- ✓ A.Apply model pruning and quantization
- B.Use a pre-trained model and fine-tune it
- C.Add more convolutional layers
- D.Increase number of filters in each layer
Why A: Model pruning removes redundant or less important weights from the CNN, reducing computational load, while quantization converts floating-point weights to lower-precision integers (e.g., INT8). Together, they shrink model size and speed up inference without significantly degrading accuracy, making them ideal for real-time object detection in resource-constrained environments like autonomous vehicles.
Last reviewed: Jun 30, 2026
This AI0-001 practice question is part of Courseiva's free CompTIA 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 AI0-001 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.