Which model specializes in grid-like data such as images and uses convolutional layers?

Which model specializes in grid-like data such as images and uses convolutional layers?

• A. Recurrent Neural Networks (RNNs)
• B. Convolutional Neural Networks (CNNs)
• C. Logistic Regression
• D. SVM

Answer: (B)

Convolutional neural networks are designed for grid-like data such as images and use convolutional layers. An image is a 2D grid of pixels, and a convolutional layer applies small filters that slide across this grid to compute local weighted sums, producing feature maps that respond to simple patterns like edges and textures. Using the same filters across the whole image means the model learns features that apply anywhere in the image, which dramatically reduces parameters and helps capture translation invariance. By stacking multiple convolutional layers, the network builds up hierarchical features—from simple edges to more complex shapes and object parts. Pooling layers then down-sample the spatial dimensions, making representations more compact and robust to small shifts. This combination is why CNNs excel with image data: they leverage spatial structure and local connectivity to learn powerful visual features. In contrast, recurrent networks are geared toward sequences, while logistic regression and SVM are general classifiers that don’t inherently exploit grid structure through convolution.

Convolutional neural networks are designed for grid-like data such as images and use convolutional layers. An image is a 2D grid of pixels, and a convolutional layer applies small filters that slide across this grid to compute local weighted sums, producing feature maps that respond to simple patterns like edges and textures. Using the same filters across the whole image means the model learns features that apply anywhere in the image, which dramatically reduces parameters and helps capture translation invariance. By stacking multiple convolutional layers, the network builds up hierarchical features—from simple edges to more complex shapes and object parts. Pooling layers then down-sample the spatial dimensions, making representations more compact and robust to small shifts. This combination is why CNNs excel with image data: they leverage spatial structure and local connectivity to learn powerful visual features. In contrast, recurrent networks are geared toward sequences, while logistic regression and SVM are general classifiers that don’t inherently exploit grid structure through convolution.