machine-learning deep-learning
Definition
Deep Artificial Neural Network
A Deep Artificial Neural Network (DANN or DNN) is an artificial neural network characterised by multiple hidden layers between the input and output layers. The “depth” of the network refers to the number of successive non-linear transformations applied to the data. Formally, a network is considered deep if it contains a sequence of hidden layers, each learning a progressively more abstract representation of the input.
Structural Properties
Feature Hierarchy: Deep networks are capable of learning hierarchical feature representations. Lower layers typically identify simple patterns (e.g., edges in an image), while deeper layers combine these into complex semantic concepts (e.g., objects or faces).
Complexity and Generalisation: Increasing the depth of a network increases its functional capacity, allowing it to solve highly non-linear tasks. However, this also increases the number of parameters, requiring larger training datasets and rigorous regularisation to prevent overfitting.