Learning Two layer Networks with Multinomial Activation and High Thresholds
arXiv: Learning, 2019.
Giving provable guarantees for learning neural networks is a core challenge of machine learning theory. Most prior work gives parameter recovery guarantees for one hidden layer networks. In this work we study a two layer network where the top node instead of a sum (one layer) is a well-behaved multivariate polynomial in all its inputs. We...More
PPT (Upload PPT)