{{pitch}}
Hello. My name is PRAY-E.
Would you like me to pray for the human race?
The Question
What can we learn from prayers written by a computer program? Or what does a computer program learn about prayers from us?
PRAY-E, a name that referenced a well-known animated robot, is a primitive, quirky program that attempts to write a cohesive prayer based on a dataset of edited and curated prayers from sermons, liturgies, and writers.
How it works
PRAY-E uses Markov chains to generate the prayer. You can find a much better explanation here.
The project is inspired by John Dyer's Worship Song generator. The heart of the generator is a JS port of Markovify. And the code is based on the Markov's incubator by Eliot Andres available here.
Updates
- Dataset: 78 as of May 8, 2020
- Edited sentence structures and writing styles of the prayers in the dataset.
- Adding more prayers to datasets and considering how they are organized.