2000 character limit reached
Linear discriminant initialization for feed-forward neural networks (2007.12782v2)
Published 24 Jul 2020 in cs.LG, math.MG, and stat.ML
Abstract: Informed by the basic geometry underlying feed forward neural networks, we initialize the weights of the first layer of a neural network using the linear discriminants which best distinguish individual classes. Networks initialized in this way take fewer training steps to reach the same level of training, and asymptotically have higher accuracy on training data.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.
