Zero Shot Learning
Hugo discusses the evolution of soft prompts and their role in transfer learning, particularly in the context of zero shot learning. He shares insights from his research on leveraging large language models for code completion, exploring how these models can utilize information from across different files in a code repository. Despite initial skepticism, he highlights the potential for these models to generalize and improve code prediction through innovative prompting techniques.In this clip
From this podcast
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Towards Improved Transfer Learning with Hugo Larochelle - 631
Related Questions
How are prompts used in AI models as discussed in the episode Treating Prompt Engineering More Like Code // Maxime Beauchemain // MLOps Podcast #167 and the clip Taming Generalized AI?
How are prompts used in AI models as discussed in the episode Treating Prompt Engineering More Like Code // Maxime Beauchemain // MLOps Podcast #167 and the clip The Future of LLMs?
How are prompts used in AI models as discussed in the episode Oriol Vinyals: Deep Learning and Artificial General Intelligence | Lex Fridman Podcast #306 and the clip Teaching Neural Networks?