Anna Garbier & Lan Zhang


An AI remake of a traditional word-guessing game.

In this new version, players explore linguistic datasets by jumping from one word to the next through constrained dialogue with a computer program. The gameplay begins when the program chooses a target word from a dataset. The player then tries to guess the target through questions and answers, beginning with “Is it more like TOAST or ___?” For example, “Is it more like TOAST or CAT?” If the target is more like CAT, then the question pivots: “Is it more like CAT or ___?” This procedure continues until the player arrives at the target.

The game's datasets were created using custom scripts for data scraping and cleaning, word2vec for converting words to word embeddings, and tsne for converting multidimensional embeddings to 2-dimensional vectors that could be mapped on an x-y plane.


Semantic mapping of datasets, before UI integration.

By Anna Garbier & Lan Zhang