The Jim Rutt Show cover image

EP 192 David Krakauer on Science, Complexity and AI

The Jim Rutt Show

NOTE

The Importance of Structure in Neural Nets

Neural nets need to have robust, devolvable, extensible, and composable key features in their structure. Internal representations of neural nets were found to have internal compositional encoding resembling schemas and constructs. Variational autoencoders are utilized to bottleneck networks for easier interrogation. Different models serve different purposes, prompting the use of a combination from various models. The ongoing development of models with specialized data sets and algorithms is leading to a diverse range of models, hinting at a potential trade-off between data set size and processing. The evolving society of minds formed by these models interacting with each other and humans is paving the way for numerous possibilities in the future.

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.
App store bannerPlay store banner