Deep Learning with R
About the training
The aim of the training is to introduce the complex deep networks while making use of R. Training materials are prepared in a way which helps understand the idea and use of deep networks especially in case of solving problems connected with learning by using data in R. Practical examples and exercises in R will show you how to create networks and tune them to a given problem.
The advantage of using the method of deep learning is the ability to predict and classify complex, non-linear problems which very often occur during the work. Microsoft, Google, IBM, Twitter, Paypal, Facebook and other big companies have already noticed the importance of using deep learning and applied it to prepare better products and services for their users. This course enables you to join the elite group of Data Scientists who use all available technologies to solve bussiness problems.
Who is this training for?
Deep Learning methods are widely used in a number of branches ranging from medicine to marketing. If you are into data analysis and use machine learning to solve problems of forecasting, classifying, finding anomalies and many others then deep learning should be a next step in your career. Data Scientists, Data Analysts and researchers who seek to increase the number of useful tools and improve the results of building models should take a closer look at deep networks.
What will I learn?
- Understand the most important elements of machine learning that are helpful in understanding the deep learning
- Learn about the idea of deep learning and understand how these networks differ from ordinary neural networks
- Understand the deep network architecture and its most important components that play a big role in data analysis
- Learn how to use deep network architectures to solve problems such as forecasting, image recognition, classification, and anomaly detection
- Learn how to build and tune deep networks using a range of R packages and an R interface for external environments
- Try various deep network applications on numerous examples during exercises in R
- Introduction to deep learning
- The concept of Machine Learning (ML)
- Why we need deep learning?
- Basic problems of ML – forecasting, classification, segmentation, anomalies detection
- Learning and validating ML algorithms
- From logistic regression to neural network
- Neural network
- Biological inspirations to Neural network
- Neural Networks– Neuron, Perceptron and MLP(Multilayer Perceptron model)
- Learning MLP – backpropagation algorithm
- Activation functions – linear, sigmoid, Tanh, Softmax
- Loss functions appropriate to forecasting and classification
- Parameters – learning rate, regularization, momentum
- Applications of Neural Networks in R
- Basics of Deep Networks
- What is deep learning?
- Architecture of Deep Networks– Parameters, Layers, Activation Functions, Loss functions, Solvers
- Restricted Boltzman Machines (RBMs)
- Deep Networks Architectures
- Deep Belief Networks(DBN) – architecture, application
- Restricted Boltzmann Machines
- Convolutional Neural Network
- Recursive Neural Network
- Recurrent Neural Network
- Overview of libraries and interfaces available in R
- Neural networks MLP- nnet, neuralnet
- Deep learning networks – deepnet, darch, rnn, autoencoder
- Interfaces to Tensorflow, keras, h20
- Building deep network
- Choosing appropriate architecture to given problem
- Hybrid deep networks
- Learning network – appropriate R package, architecture definition
- Tuning network – initialization, activation functions, loss functions, optimization method
- Avoiding overfitting – detecting overfitting problems in deep networks, regularization
- Deep learning – solving common problems – with applications in R
- Image recognition – CNN
- Detecting anomalies with Autoencoders
- Forecasting time series with RNN
- Dimensionality reduction with Autoencoder
- Classification with RBM