AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Is Memory Differentiable?
Oftentimes, there are ideas that we might want to apply in the contexts of neural networks. The big challenge is, yet, are they differentiable, so that we can apply techniques like gradient dissent? How does the size of the memory playinto whether we can back prop over it or not? Are there ways that can be harnessed more directly to achieve some of your galls for episodic memory?