A Hybrid Architecture Makes Convex Neural Networks Actually Scale — and the Proof Is Exponential Parameter Savings

A Hybrid Architecture Makes Convex Neural Networks Actually Scale — and the Proof Is Exponential Parameter Savings | BedrockNews