Learning on a Razor's Edge: Identifiability and Singularity of Polynomial Neural Networks

arXiv:2505.11846v2 Announce Type: replace-cross Abstract: We study function spaces parametrized by neural networks, referred to as neuromanifolds. Specifically, we focus on deep Multi-Layer Perceptrons (MLPs) and Convolutional Neural Networks (CNNs) with an activation function...

🔗 Read more: https://arxiv.org/abs/2505.11846

#News #Neuro #Math #Policy #AI #Biology #Academic