Skip to main content
Browse by:
GROUP

Humanities +DS and Social Sciences +DS: Text Generation with GPT-2

Event Image
Icon calendar
Wednesday, September 18, 2019
Icon time
4:00 pm - 6:00 pm
Icon speaker
Matthew Kenney, MFA
Icon series
+DS In-Person Learning Experiences

With recent advancements in natural language generation, AI models can now produce highly coherent generative text. In this IPLE, we will focus on creative and practical applications for language generation. First, we will start by overviewing current state-of-the-art AI models for text generation, as well as practical use cases and creative applications. Next, we will inspect the architecture of models like GPT-2 to understand how generative text models work. Lastly, we will take a hands-on approach, fine-tuning GPT-2 to generate poetry, stories, or conversation.

This session is part of the Duke + Data Science (+DS) program. To learn more, please visit https://plus.datascience.duke.edu/