Download: HTTP: // gqylpy / DI / 199
"Deep learning, optimization and identification Jiao Licheng (detail bookmark) .pdf." PDF HD full version - Download
The depth of neural network research in recent years by the widespread attention, it has become a major component of artificial intelligence 2.0. "Deep learning, optimization and recognition" systematically discusses the basic theory of deep neural networks, algorithms and applications. "Deep learning optimization and recognition" book chapters 16, divided into two parts; the first part (Chapter 1 - Chapter 10) discusses the theoretical system and algorithm, feedforward neural network comprising a depth that convolutional neural network, a depth stack neural network, recurrent neural network depth, depth generation network, the depth of integration networks; the second part (Chapter 11 - 15) discusses the common depth of learning platforms, and hyperspectral image, natural image, polarization SAR and SAR in video and other applications; Chapter 16 for the summary and outlook, given the history of depth learning and development of cutting-edge direction and progress. "Deep learning, optimization and recognition" Each chapter is accompanied by the relevant reading material and simulation code so that the interested reader to delve into further.
"Deep learning, optimization and recognition" to institutions of higher learning computer science, electronic science and technology, information science, control science and engineering fields, artificial intelligence researchers to provide a reference, as well as related professional undergraduate and graduate teaching reference books while learning applications for its depth of interest of researchers and engineers and technical officers.
table of Contents
Chapter 1 deep learning foundation
1.1 mathematical basis
1.1.1 Matrix Theory
1.1.2 Probability
1.1.3 Optimization
1.1.4 Frame Analysis
1.2 Sparse Representation
1.2.1 sparse representation preliminary
1.2.2 sparse model
1.2.3 sparse cognitive learning, and recognition calculation paradigms
1.3 machine learning and neural networks
1.3.1 Machine Learning
1.3.2 Neural Networks
references
Chapter 2 depth feedforward neural networks
2.1 biological mechanism neurons
2.1.1 Biological Mechanism
2.1.2 single hidden layer feed forward neural network
More than 2.2 before the hidden layer feedforward neural network
2.3 back-propagation algorithm
2.4 feedforward neural network of the former deep learning paradigm
references
Chapter 3, depth convolution neural network
Biological mechanism 3.1 convolution neural networks and mathematical characterization
3.1.1 Biological Mechanism
3.1.2 mathematical characterization of convolution stream
3.2 depth of convolution neural network
3.2.1 typical network models and frameworks
3.2.2 learning algorithm and Training Strategy
Advantages and disadvantages of the model 3.2.3
3.3 depth deconvolution neural network
3.3.1 convolution sparse coding
3.3.2 depth deconvolution neural network
Performance Analysis and Application Examples 3.3.3 Network Model
3.4 full convolution neural network
3.4.1 describe the mathematical model of the network
Performance Analysis and Application Examples 3.4.2 Network Model
references
Chapter 4 from the stack depth coding network
4.1 Since the network coding
4.1.1 step by step learning strategies
4.1.2 self-encoded network
4.1.3 common paradigm from coding network
4.2 network stack depth
4.3 belief networks depth / depth Boltzmann machine network
4.3.1 Boltzmann machine / Restricted Boltzmann Machine
4.3.2 Boltzmann machine depth / depth belief networks
references
Chapter 5 sparse depth neural network
5.1 biological mechanism sparsity
5.1.1 Mechanism of biological vision
5.1.2 sparsity physical and mathematical description of the response
5.2 Sparse depth network model and basic properties
5.2.1 Data sparsity
5.2.2 sparse regularization