Natural language processing
These images were created with a machine learning technique called word embeddings. A computational model learns the meaning of a word based on the other words surrounding it in context. These embeddings are unique representations of a word, defined by context. Embeddings can be decomposed into a vector, which allows a visual representation of each poem. In these works, every shape outline is a word in the respective poem. Its position, rotation, and size are determined by the word embedding in vector form.