About this document
Optimizing Multilayer Perceptron Architecture by Aquacypress is a document available to read on EtoBox.
1. The document discusses optimizing the architecture of multilayer perceptron (MLP) neural networks, specifically determining the optimal number of hidden layers and neurons in each layer. 2. It reviews previous approaches to neural network architecture optimization and notes limitations like fixing the architecture before training or not finding truly optimal architectures. 3. The authors propose modeling the architecture optimization problem as a mixed-integer nonlinear problem to determine a network
- Author
- Aquacypress
- Language
- EN