Convolutional Neural Network (CNN)

Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary

A neural network built for grid-shaped data such as images, using convolutional filters that slide across the input to learn spatial patterns layer by layer.

What is a Convolutional Neural Network (CNN)?

A Convolutional Neural Network (CNN) is a deep learning model built for grid-shaped data such as images, using filters that slide across the input to detect local patterns like edges and textures.

The trick is weight sharing: instead of learning a separate parameter for every pixel, a CNN reuses one small filter across the whole image. A cat detector learned in the top-left corner works in the bottom-right too, which keeps the network small enough to train on real photos.

How a CNN Works

  1. Convolutional layers slide filters over the input and produce feature maps. Early layers respond to edges and color blobs; deeper layers combine those into shapes, parts, and objects.
  2. Pooling layers shrink each feature map, keeping the strongest signals while cutting computation and making the network less sensitive to small shifts in position.
  3. Fully connected layers sit at the end and turn the extracted features into a final prediction, such as a class label.

CNN vs RNN

The practical difference: a CNN reads spatial data all at once, while a recurrent neural network reads sequential data one step at a time. A CNN asks “which pixels sit next to which”; an RNN asks “what came before this”.

CriterionCNNRNN
Data shapeGrid-like: images, video frames, spectrogramsSequential: text, audio, time series
What it exploitsSpatial locality between neighboring valuesOrder and context from earlier steps
ProcessingWhole input in parallelStep by step, carrying a hidden state
Typical tasksImage classification, object detectionLanguage modeling, forecasting

Applications of CNNs

Example of a Convolutional Neural Network

AlexNet, a CNN trained by Alex Krizhevsky and colleagues, won the 2012 ImageNet image-classification competition by a wide margin over non-neural methods. That result convinced much of the field that deep learning was the path forward for computer vision.

Follow a photo through it: convolutional filters first fire on edges and textures, deeper layers assemble those into whiskers and ears, pooling discards position noise along the way, and the final layers output “cat” with a confidence score.

Related AI terms: Recurrent Neural Network · Model Architecture · Deep Learning · Computer Vision · Image Recognition

Did you like the Convolutional Neural Network (CNN) gist?

Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.

Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About

Read the Governor's Letter

Stay ahead with Governor's Letter, the newsletter delivering expert insights, AI updates, and curated knowledge directly to your inbox.

By subscribing to the Governor's Letter, you consent to receive emails from AI Guv.
We respect your privacy - read our Privacy Policy to learn how we protect your information.

Browse All AI Terms A–Z

Every term in the dictionary, in alphabetical order. Jump to a letter or scroll the full list.

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z