Deep learning aims to bring machine learning closer to one of its original goals: Artificial Intelligence (AI). AI encompasses a wide range of technologies that enable computers to solve problems in a way that (at least superficially) resembles human thinking. Within that realm is a smaller category called machine learning, which is the name for a whole toolbox of mathematical techniques that allow computers to improve their performance by performing tasks. Finally, within machine learning, there is the smaller subcategory of deep learning. Deep learning focuses on a subset of machine learning tools and techniques through the use of deep neural networks.
Deep neural networks are distinguished from other neural networks by their depth. A deep neural network is a neural network with multiple hidden layers between the input and output layers. The large number of layers makes complex transformations possible. In short, deep learning is the application of (deep) neural networks that contain more than one hidden layer.
Research into neural networks made the term deep learning popular because it could be emphasized that the researchers were able to train deeper neural networks and draw attention to the theoretical importance of this depth. Many scientists, therefore, prefer to use the academic term deep neural networks rather than deep learning.
Above all, these deep neural networks are capable of discovering latent structures within unstructured and unlabeled data, or the vast majority of all data in the world. One of the problems that deep learning excels in is the processing and bundling of this raw data. In other words: To distinguish patterns in data that no person has ever organized or given a name.
Deep learning techniques are currently the state of the art for identifying pattern recognition, such as objects in images and words in sounds. Deep learning enables computers to learn representations of data with multiple levels of abstraction: From properties at the lowest level to concepts at the highest level.
