Organizing and representing a collection of fonts according to visual similarity utilizing machine learning

A system comprising:

• one or more memory devices comprising a plurality of fonts, an encoder neural network, and a self-organizing map;
• and
• one or more computing devices that are configured to cause the system to:
• extract feature vectors from the plurality of fonts utilizing the encoder neural network;
• generate, utilizing the self-organizing map, a font map comprising a set of nodes corresponding to the plurality of fonts by:
• determining feature distances between the feature vectors and node weights corresponding to the set of nodes;
• and
• mapping particular feature vectors representing visually similar fonts to corresponding nodes within the set of nodes based on the feature distances;
• and
• provide, for display on a client device, a visual depiction of one or more fonts from the plurality of fonts based on the font map.

A computer-implemented method for generating and providing font maps for font selection, the computer-implemented method comprising:

• identifying a plurality of fonts of varying styles associated with a client device;
• performing a step for constructing a font map arranging the plurality of fonts according to visual similarity;
• and
• providing, for display on the client device, a graphical user interface comprising visual depictions of fonts from the plurality of fonts organized for navigation according to the font map.