Название: How to Fine-tune Neural Networks for Classification
Автор: Ionut Brandusoiu, Gavril Toderean
Издательство: GAER Publishing House
Год: 2020
Язык: английский
Формат: pdf (true)
Размер: 14.9 MB

This book covers in the first part the theoretical aspects of neural networks and their functionality, and then based on the discussed concepts it explains how to find-tune a neural network to yield highly accurate prediction results which are adaptable to any classification tasks. The introductory part is extremely beneficial to someone new to learning neural networks, while the more advanced notions are useful for everyone who wants to understand the mathematics behind neural networks and how to find-tune them in order to generate the best predictive performance of a certain classification model.

For a better understanding of the underlying theory of neural networks for classification, Chapter 1 describes supervised learning in the context of machine learning and all its related concepts. In terms of the generalization error as the metric to evaluate the performance of a classification model, the second half of this chapter contrasts between theoretical and empirical estimations.

Chapter 2 starts with the introduction to neural networks, followed by theoretical foundations pertaining to the multilayer perceptron and the backpropagation learning algorithm. It discusses several noteworthy research papers existent in the literature regarding techniques for optimizing the structure of neural networks and accelerating the learning process. It introduces the simulated annealing algorithm for weights initialization.