Artificial Intelligence By Example

Dimensionality reduction

Pert is in trouble, and he knows it. Not all the data Pert would have liked can be obtained for the moment. Some features must be left aside for the moment. He thinks back to his walk from the IT manager's office to his office and looks at his cup of coffee.

Dimensionality reduction comes up!

DL uses dimensionality reduction to reduce the number of features in, for example, an image. Each pixel of a 3D image, for example, is linked to a neuron, which in turn brings the representation down to a 2D view with some form of function. For example, converting a color image into shades of a now-gray image can do the trick. Once that is done, simply reducing the values to, for example, 1 (light) or 0 (dark) makes it even easier ...