Machine Learning with TensorFlow

In the course Machine Learning with TensorFlow from SpiralTrain participants learn to implement machine learning and deep learning applications with the open source TensorFlow framework.

TensorFlow Machine Learning

The course Machine Learning with TensorFlow starts with an overview of the basic principles of Machine Learning and an explanation of the differences of Supervised, Unsupervised and Deep Learning. The data types of TensorFlow like vectors, arrays, lists and scalars are treated and the Colab and DataBricks development environments are discussed.

Tensors

Subsequently the Machine Learning with TensorFlow course pays attention to the central Tensor Data Structure, which can be regarded as a container in which data in N dimensions can be stored. Rank, shape and type of tensors are discussed and TensorFlow operations and sessions are also treated.

Neural Networks

Special attention is given to neural networks in which both Convolutional and Recurrent Neural Networks are explained. Convolution and Pooling, making connections between Input Neurons and Hidden Layers are also discussed.

Model Visualization

The Visualization of models with TensorBoard is also part of the Machine Learning with TensorFlow course. Supervised Learning with Linear and Logistic Regression are reviewed and Ensemble techniques and Gradient Boosting are explained.

Text Processing

In addition the course Machine Learning with TensorFlow deals with Natural Language Processing with tokenization and text classification. Spam detection serves as an example and also Deep Learning is on the course schedule.

TensorFlow Optimizers

Various TensorFlow Optimizers such as Stochastic Gradient Descent, Gradient clipping and Momentum are discussed as well. And also Image Processing with Dimensionality Reduction and using the Keras APIs is covered.

Model Deployment

Finally the course Machine Learning with TensorFlow ends with a discussion of models in production. Models as REST Service and Keras Based Models are treated.

Audience Course Machine Learning with Tensor Flow

The course Machine Learning with TensorFlow is intended for data scientists who want to use Python and the TensorFlow machine learning libraries to make predictions based on models.

Prerequisites for course Machine Learning with TensorFlow

To participate in this course knowledge of and experience with Python is required and knowledge of data analysis libraries such as Numpy, Pandas and Matplotlib is desirable.

Realization training Machine Learning with TensorFlow

The theory is discussed on the basis of presentations. Illustrative demos clarify the concepts. The theory is interchanged with exercises. The Anaconda distribution with Jupyter notebooks is used as a development environment. Course times are from 9:30 to 16:30.

Official Certificate Machine Learning with TensorFlow

After successful completion of the course participants receive an official certificate Machine Learning with TensorFlow.

Modules

Module 1 : Intro TensorFlow

• What is TensorFlow?
• Machine Learning
• Supervised Learning
• Unsupervised Learning
• Deep Learning
• Install Anaconda
• Install TensorFlow
• Colab and Databricks
• Vectors and Scalars
• Matrix Calculations

Module 2 : Tensor Data Structure

• Arrays and Lists
• Multiple Dimensions
• Rank, Shape and Type
• TensorFlow Dimensions
• Tensor Manipulations
• TensorFlow Graphs
• Variables and Constants
• TensorFlow Operations
• TensorFlow Sessions
• Placeholders

Module 3 : Neural Networks

• What are Neural Networks?
• Convolutional Neural Networks
• Multiple Layers of Arrays
• Local respective fields
• Convolution and Pooling
• Connecting Input Neurons
• Hidden Layers
• Recurrent Neural Networks
• Sequential Approach
• Layer Independence

Module 4 : Tensor Board

• Data Visualization
• Data Flow Graph
• High Level Blocks
• High Degree Nodes
• Node Representations
• Sequence Numbered Nodes
• Connected Nodes
• Operation Nodes
• Summary Nodes
• Reference Edge

Module 5 : Supervised Learning

• Linear Regression
• Keras and TensorFlow
• Correlation Graph
• Pairplot
• Logistic Regression
• Categorical Outcomes
• Sigmoid Function
• Boosted Trees
• Ensemble Technique
• Gradient Boosting

Module 6 : Natural Language Processing

• NLP Overview
• NLP Curves
• Text Preprocessing
• Tokenization
• Spam Detection
• Word Embeddings
• Deep Learning Model
• Text Classification
• Text Processing
• TensorFlow Projector

Module 7 : TensorFlow Optimizers

• Stochastic Gradient Descent
• Gradient clipping
• Momentum
• Nesterov momentum
• Adagrad
• Adadelta
• RMSProp
• Adam
• Adamax
• SMORMS3

Module 8 : Image Processing

• Convolution Layer
• Pooling Layer
• Fully Connected Layer
• Keras API's
• ConvNets
• Transfer Learning
• Autoencoders
• Dimensionality Reduction
• Compression Techniques
• Variational Autoencoders

Module 9 : Models in Production

• Model Deployment
• Isolation
• Collaboration
• Model Updates
• Model Performance
• Load Balancer
• Model as REST Service
• Templates
• Keras Based Models
• Flask Challenges