Natural Graph Networks with Taco Cohen - #440

This article summarizes a podcast episode of Practical AI featuring Taco Cohen, a machine learning researcher. The discussion centers around Cohen's research on equivariant networks, video compression using generative models, and his paper on "Natural Graph Networks." The paper explores "naturality," a generalization of equivariance, suggesting that less restrictive constraints can lead to more diverse architectures. The episode also touches upon Cohen's work on neural compression and a visual demonstration of equivariant CNNs. The article provides a brief overview of the topics discussed, highlighting the key research areas and the potential impact of Cohen's work.