Quotient Geometry, Effective Curvature, and Implicit Bias in Simple Shallow Neural Networks

Quotient Geometry, Effective Curvature, and Implicit Bias in Simple Shallow Neural Networks explores Develops a differential-geometric framework to analyze shallow neural networks by accounting for parameter symmetries and their intrinsic geometric properties.. Commercial viability score: 2/10 in Neural Network Theory.