Skip to main content

Convolutional Neural Network

What is Convolutional Neural Network​

A Convolutional Neural Network (ConvNet/CNN) is a Deep Learning algorithm which can take in an input image, assign importance (learnable weights and biases) to various aspects/objects in the image and be able to differentiate one from the other. The pre-processing required in a ConvNet is much lower as compared to other classification algorithms. While in primitive methods filters are hand-engineered, with enough training, ConvNets have the ability to learn these filters/characteristics.

Convolutional Neural Network using PyTorch​

Import the necessary packages for creating a simple neural network.

from torch.autograd import Variable
import torch.nn.functional as F

Create our simple convolutional neural network class

class SimpleCNN(torch.nn.Module):
    def __init__(self):
        # define the structure of our network
    super(SimpleCNN, self).__init__()
    #Input channels = 3, output channels = 18
    self.conv1 = torch.nn.Conv2d(3, 18, kernel_size = 3, stride = 1, padding = 1)
    self.pool = torch.nn.MaxPool2d(kernel_size = 2, stride = 2, padding = 0)
    #4608 input features, 64 output features (see sizing flow below)
    self.fc1 = torch.nn.Linear(18 * 16 * 16, 64)
    #64 input features, 10 output features for our 10 defined classes
    self.fc2 = torch.nn.Linear(64, 10)
    def forward(self, x):
      x = F.relu(self.conv1(x))
        x = self.pool(x)
      x = x.view(-1, 18 * 16 *16)
      x = F.relu(self.fc1(x))
      #Computes the second fully connected layer (activation applied later)
      #Size changes from (1, 64) to (1, 10)
      x = self.fc2(x)
      return(x)

Feature Visualization on Convolutional Neural Network​

Deep Neural Networks are usually treated like “black boxes” due to their inscrutability compared to more transparent models, like XGboost or Explainable Boosted Machines.

However, there is a way to interpret what each individual filter is doing in a Convolutional Neural Network, and which kinds of images it is learning to detect by using Feature Visualization.

Here are additional resources on convolutional networks and feature visualization.

Acknowledgement : The content of this document has been adapted from these original websites.