
31 - Singular Learning Theory with Daniel Murfet
AXRP - the AI X-risk Research Podcast
Unraveling Neural Network Dynamics
This chapter explores the relationship between parameter space symmetries and degeneracies in neural networks, emphasizing their influence on learning behavior. It specifically discusses the local learning coefficient within the context of singular learning theory, illustrating its impact on model accuracy and complexity. Various examples are provided to deepen understanding of how these concepts shape the dynamics of deep learning and Bayesian inference.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.