Deep learning (also known as deep structured learning or hierarchical learning) is part of a larger group of machine learning approaches based on learning data representations, as opposed to task-specific algorithms. 

Learning can be supervised (which we covered in Chapter 3, Supervised Machine Learning Models and Your Data), semi-supervised, or unsupervised (covered in Chapter 4, Implementing Unsupervised Algorithms).

Deep learning algorithms are at work in exciting areas such as image classification (categorizing every pixels in a digital image into one of several land cover classes, or themes), object detection (the process of finding instances of real-world objects such as faces, cars, and buildings in images or videos), image restoration (to compensate for, or undo, defects caused by motion blur, noise, and camera misfocus, which degrade an image) and image segmentation (the process of partitioning a digital image into multiple segments of pixels, also known as super-pixels, to simplify and/or change the representation of an image into something that is more meaningful and easier to analyze). 

Deep learning using enormous neural networks is teaching machines to automate the tasks performed by human visual systems. 

Deep learning models are vaguely inspired by information processing and communication patterns in biological nervous systems, yet they do differ from the structural and functional properties of biological brains, which make them incompatible with neuroscience evidences.

Enough theory. While the preceding explanation of machine/deep learning may be high-level, it is sufficient enough for us to move on to the next section, where we start thinking about the means of deep learning implementation, specifically using a toolset developed by the Google Brain team for internal Google use, under the Apache 2.0 open source license on November 9, 2015: TensorFlow.