Skip to content Skip to navigation

Generating Homeric Poetry with Deep Neural Networks

Transdisciplinary AI cover
America Chambers
2019 First International Conference on ​Transdisciplinary AI
May 2020

We investigate the generation of metrically accurate Homeric poetry. We assess the quality of the generated lines of poetry using quantitative metrical analysis and expert evaluation. This evaluation reveals that our model captures complex poetic meter implicitly, but underperforms in terms of semantics and context matching. Our research highlights the importance of expert evaluation and shows that future research should focus on poetic semantics.