LSTM-based poetry generator

Client: Experimenta Heilbronn
Agency: Natural Vision

An interactive museum exhibit to create one’s own poem with the help of Artificial Intelligence.

Goal:
To integrate an LSTM algorithm with an intuitive interface.

Contribution:
I implemented a full-stack solution to interact with a pre-existing LSTM library using PyTorch and Flask, along with a fully custom WebGL-enabled front-end, based on PixiJS.

Media

Exhibit installation
Exhibit installation
Exhibit interaction
Exhibit interaction