1 paper
Paul Lezeau, Thomas Walker, Yueqi Cao +2
We propose an algebraic geometric framework to study the expressivity of linear activation neural networks. A particular quantity of neural networks that has been actively studied…