Learning to Ponder: Memory in Deep Neural Networks with Andrea Banino - #528

This article summarizes a podcast episode featuring Andrea Banino, a research scientist at DeepMind. The discussion centers on artificial general intelligence (AGI), specifically exploring episodic memory within neural networks. The conversation delves into the relationship between memory and intelligence, the difficulties of implementing memory in neural networks, and strategies for improving generalization. A key focus is Banino's work on PonderNet, a neural network designed to dynamically allocate computational resources based on problem complexity. The episode promises insights into the motivations behind this research and its connection to memory research.