BDP (Big-Data Poetry) currently investigates machine learning and neural networks as tools for literary creation. The work is ongoing and open-ended.
2016-2017 BDP: investigates machine learning using Tensorflow, Keras and PyTorch. Neural net models are trained on a custom corpus of 600,000 lines of contemporary poetry: from the romantic epoch to the 20th century avant garde. Successful models are then sequentially asked to generate poems in an infinite loop.
The live poem output is now used as a projection in conjunction with spoken word performances. The Python-based terminal then becomes a site for writing by the machine, reading by the human poet who must attempt to stitch and weave together poems from the incoherent hybrid yet often astonishing word-debris.
Current installation mode can be for as many screens as are available in a space. The generation process runs easily on laptop. Future iterations will involve real-time synthetic audio during performances to accompany spoken word (controlled by Leap and a neural net inside Wekinator).
Previously in 2011-2014, BDP used a combination of data visualization, language analytics, classification algorithms, entity recognition and part-of-speech replacement techniques. Based on these templates, a Python script generated thousands of poems per hour. Sometimes Jhave reads along with this writing machine, verbally stitching and improvising spoken poems.
Jhave is a digital poet, prof , & author of Aesthetic Animism: Digital Poetry’s Ontological Origins (MIT Press, 2016).