Episode · Jun 4, 2025
Grokking, Generalization Collapse, and the Dynamics of Training Deep Neural Networks
The TWIML AI Podcast
We use cookies for analytics.
Podcast, The TWIML AI Podcast
Hosted by Sam Charrington
Andrea Banino discusses episodic memory in artificial neural networks and how the ability to store and use prior experiences relates to generalization. The episode also covers models that learn how long to compute before making a prediction and the challenges of developing flexible memory for learning systems.