Deep Learning with C#, .Net and Kelp.Net

Name: Deep Learning with C#, .Net and Kelp.Net | The Ultimate Kelp.Net Deep Learning Guide
Brand: De Gruyter
Price: 14.49 EUR
Availability: OnlineOnly

The Ultimate Kelp.Net Deep Learning Guide

Matt R. Cole(Autor*in)

De Gruyter (Verlag)

1. Auflage

Erschienen am 15. Januar 2019

412 Seiten

E-Book

ePUB mit Adobe-DRM

Systemvoraussetzungen

978-93-89423-74-7 (ISBN)

14,49 €inkl. 7% MwSt.

Systemvoraussetzungen

für ePUB mit Adobe-DRM

E-Book Einzellizenz

Als Download verfügbar

Beschreibung

Weitere Details

Weitere Ausgaben

Inhalt

Cover
Deep Learning with C#, .NET and Kelp.NET
Copyright
About the Author
Reviewer
Preface
Acknowledgement
Errata
Table of Contents
1. Take This ___ and ___ It
Objectives of this book
Neural network overview
Machine learning overview
Deep learning overview
Complexity
Machine and deep learning differences
Summary
2. Machine Learning/Deep Learning Terms and Concepts
Overview
Neuron/Perceptron
Multi-Layer Perceptron (MLP)
Features
Weights
Bias
Activation Function
Sigmoid
ReLU (Rectified Linear Units)
Softmax
Neural network
Input/Output/Hidden Layers
Forward propagation
Back propagation
The No Free Lunch theorem
The Curse of Dimensionality
The more neurons versus more layers
Cost function
Gradient descent
Learning rate
Batches/Batch size
Epochs
Iterations
Dropout
Batch Normalization
CNN (Convolutional Neural Network)
Pooling
Padding
Recurrent neuron
RNN (Recurrent Neural Network)
Vanishing gradient problem
Exploding gradient problem
Logistic Neurons
Hidden layers
Types of neural networks
Generalization
Regularization
Loss
Loss over time
Loss versus learning curve
Supervised learning
Bias-Variance Trade-off (overfitting and underfitting)
Bias
Variance
Overfitting
Is your model overfitting or underfitting?
Prevention of overfitting and underfitting
Amount of training data
Input space dimensionality
Incorrect output values
Data heterogeneity
Unsupervised learning
Reinforcement learning
Manifold learning
Types of manifolds in deep learning
Topological
Differentiable
Riemannian
Principal Component Analysis (PCA)
Hyperparameter training
Approaches to hyperparameter tuning
Grid search
Random search
Bayesian optimization
Gradient-based optimization
Evolutionary optimization
Summary
References
3. Deep Instrumentation Using ReflectInsight
Next generation logging viewers
Message log
Message details
Message properties
Bookmarks
Call Stack
Message Navigation
Advanced Search
User-Defined Views and Filtering
Auto Save/Purge rolling log files
Watches
Time zone formatting
Router
Log viewer
Live viewer
SDK
Configuration editor
Overview
XML configuration
Dynamic configuration
Configuration editor
Message type logging reference
Assertions
Assigned variables
Attachments
Audit failure and success
Checkmarks
Checkpoints
Collections
Comments
Currency
Data
DataSet
DataSetSchema
DataTable
DataTableSchema
DataView
Date/Time
Debug
Desktop Image
Errors
Exceptions
Fatal Errors
Generations
Images
Information
Levels
Linq queries and results
Loaded assemblies
Loaded processes
Memory status
Messages
Notes
Process Information
Reminders
Serialized Objects
SQL strings
Stack Traces
System Information
Text files
Thread Information
Typed collections
Warning
XML
XML files
Tracing method calls
Attaching message properties
To one request
To all requests
To a single message
Watches
Using custom data
Output
Summary
4. Kelp.Net Reference
Let us be honest
Downloading Kelp.Net
Building the source code
What is Kelp.Net?
N-dimensional arrays
Optimizers
AdaDelta
AdaGrad
Adam
GradientClippin g
MomentumSGD
RMSprop
SGD
Poolings
MaxPooling
AveragePooling
FunctionStack
FunctionDictionary
SplitFunction
SortedList
SortedFunctionStack
Activation Functions
Activation plots
ArcSinH
ArcTan
ELU
Gaussian
LeakyReLU
LeakyReLUShifted
LogisticFunction
MaxMinusOne
PolynomialApproximantSteep
QuadraticSigmoid
RbfGaussian
ReLU
ReLuTanh
ScaledELU
Sigmoid
Sine
Softmax
Softplus
SReLU
SReLUShifted
Swish
Tanh
Connections
Convolution2D
Deconvolution2D
EmbedID
Linear
LSTM
Normalization
BatchNormalization
Local Response Normalization
Noise
Dropout
StochasticDepth
Loss
MeanSquaredError
SoftmaxCrossEntropy
Datasets
CIFAR-10
CIFAR-100
MNIST
Street View House Numbers (SVHN)
Summary
References
5. Model Testing and Training
Accuracy
Timing
Common stacks
Summary
6. Loading and Saving Models
Loading models
Saving models
Model size
Summary
7. Sample Deep Learning Tests
A simple XOR problem
Complete source code
Output
A penny for your thoughts
A simple XOR problem (part 2)
Complete source code
Output
Recurrent Neural Network Language Models (RNNLM)
Complete source code
Vocabulary
Output
Word prediction test
Complete source code
Output
Decoupled Neural Interfaces using Synthetic Gradients
Output
MNIST accuracy tester
Complete source code
Output
Massively Deep Network Test
Complete source code
Output
Image prediction test
Complete source code
Output
Function benchmarking
Output
MNIST (handwritten characters) learning test
Complete source code
Output
LeakyReLu and PolynomialApproximantSteep Combination Network
Complete source code
Output
FunctionStack navigation tests
Complete source code
Output
Learning Rate Hyperparameter tester
Complete source code
Output
Model scoring
Complete source code
Output
Summary
8. Creating Your Own Deep Learning Tests
Example
Implementing the Run function
Create a FunctionStack with your functions
Set the optimizer
Make your predictions
Save the model
Loading models
Summary
Thank You
Appendix A
Evaluation metrics
Metrics terminology
Confusion matrix
Appendix B
OpenCL
OpenCL hierarchy

Systemvoraussetzungen

Als PDF speichern Als Link merken