Stateful LSTM, CharRNNs, and ML5.js Memo Akten Artist Residency

Memo Akten Artist Residency (2019)

The computational artist, Memo Akten, explores the interactions and tensions between ecology, technology, science, and spirituality. During his 2019 residency in the Studio, he contributed to ML5.js, an open-source software toolkit that provides artists with resources and knowledge for implementing machine learning in their art and designs. It makes machine learning accessible to a wide range of audiences, more than visual artists but also to creative coders and students. 

Memo Akten developed the Stateful LSTM and CharRNNs toolkit in ML5. js. As a type of Neural Network architecture, RNN and LSTMs (Long Short Term Memory networks) are useful for working with sequential data (like characters in text or the musical notes of a song) where the order of the sequence matters.

The ML5.js website has a class that allows users to run a pre-trained model on a body of text for generating new text. Other handy tutorials and machine-learning models of images, sounds, and texts can be found on their website.

Check them out