What Does Deep Learning Mean?
Deep learning is an iterative approach to artificial intelligence (AI) that stacks machine learning algorithms in a hierarchy of increasing complexity and abstraction. Each deep learning level is created with knowledge gained from the preceding layer of the hierarchy.
The first layer of a deep image recognition algorithm, for example, might focus on learning about color patterns in training data, while the next layer focuses on shapes. Eventually, the hierarchy will have layers that focuses on various combinations of colors and shapes, with the top layer focusing on the actual object being recognized.
Deep learning is currently the most sophisticated AI architecture in use today. Popular deep learning algorithms include:
Convolutional neural network – the algorithm can assign weights and biases to different objects in an image and differentiate one object in the image from another. Used for object detection and image classification.
Recurrent neural networks – the algorithm is able to remember sequential data. Used for speech recognition, voice recognition, time series prediction and natural language processing.
Long short-term memory networks – the algorithm can learn order dependence in sequence prediction problems. Used in machine translation and language modeling.
Generative adversarial networks – two algorithms compete against each other and use each other’s mistakes as new training data. Used in digital photo restoration and deepfake video.
Deep belief networks – an unsupervised deep learning algorithm in which each layer has two purposes: it functions as a hidden layer for what came before and a visible layer for what comes next. Used in healthcare sectors for cancer and other disease detection.
Techopedia Explains Deep Learning
Deep learning is used to build and train neural networks and decision-making network nodes. It is considered to be a core technology of the Fourth Industrial Revolution (Industry 4.0) and Web3.
Deep learning removes the manual identification of features in data and, instead, relies on whatever training process it has in order to discover the useful patterns in the input examples. This makes training the neural network easier and faster, and it can yield a better result that advances the field of artificial intelligence.
An algorithm is considered to be deep if the input data is passed through a series of nonlinearities or nonlinear transformations before it becomes output. Today, most business applications use shallow machine learning algorithms.
Shallow AI, also referred to as narrow AI, does not build a hierarchy of subroutine calls. Instead, this type of learning algorithm is designed to perform a single, discrete task.