Language Model Challenges
Hilary discusses the inherent issues with pre-trained language models, emphasizing the challenges posed by unfiltered data from the internet. She explains their approach to language generation, which involves using curated templates to ensure quality and control, while acknowledging the trade-offs in generalizability. The conversation also touches on the evolution of data science practices, highlighting the importance of structured processes in tackling complex problems.In this clip
From this podcast
Super Data Science: ML & AI Podcast with Jon Krohn
SDS 589: Narrative A.I. — with Hilary Mason
Related Questions