Open shiffman opened 6 years ago
https://github.com/shiffman/p5-word2vec https://gist.github.com/aparrish/2f562e3737544cf29aaf1af30362f469
What about using word embeddings to make a Semantris clone: https://research.google.com/semantris/ (Thinking mostly of the Blocks game). Cosine similarity between words can be used to find the "nearest" block.
https://github.com/shiffman/p5-word2vec https://gist.github.com/aparrish/2f562e3737544cf29aaf1af30362f469