The relationships between words can be derived by looking at their relative placement with respect to each other. These relationships can be viewed as a time series wherein words that are spoken can be thought of as constituting a time series database. On the other hand, we can view their relative positions and derive relationships out of these. These approaches are used by more complex and modern forms of Artificial Neural Networks (ANNs), known as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). Here, we will deep dive into CNNs and understand how they help us solve problems for the textual domain.
We will begin by understanding what a CNN is and view the various components in the CNN architecture. We will try and form an understanding of convolutions as an operation, followed by exploring the various layers that comprise...
