Contrast CNNs and fully connected networks
Company: Amazon
Role: Applied Scientist
Category: Machine Learning
Difficulty: easy
Interview Round: Technical Screen
Compare **convolutional neural networks (CNNs)** with **fully connected (dense) networks**.
Explain:
- The structural differences between convolutional layers and fully connected layers (parameter sharing, locality, connectivity pattern).
- Why CNNs are more efficient for certain data types.
- What kinds of **data or tasks** are most suitable for CNNs, and when a fully connected architecture is more appropriate.
Overview: This question evaluates understanding of neural network architectures, contrasting convolutional neural networks and fully connected (dense) networks by focusing on parameter sharing, locality, and connectivity patterns in the Machine Learning domain.
