LEARN TENSORFLOW.JS TODAY
Machine Learning is transforming our world. In recent years, advances in Artificial Neural Network algorithms and library implementations have made it possible to utilise the power of Deep Learning in commercial applications. Now is the time to get on board with the Machine Learning Revolution!
In this course we will show you how to master TensorFlow.js. We have packed hundreds of hours of learning and experimentation into 6 hours of video. We’ve then added quizzes and programming labs which you can use to solidify your knowledge.
HI, I’M JUSTIN EMERY
“I wanted to create a course that explained machine learning in a more accessible way, using visualisations instead of complex math. In this course I balance practical experience with TensorFlow.js, and the conceptual understanding of machine learning and neural networks.”
He graduated with 1st class honours from The University of Manchester in BSc Computer Science. Later, whilst working full time, he studied at Birkbeck, University of London, gaining an MSc Advanced Information Systems.
“In 2010, during my MSc studies, I was introduced to machine learning and neural networks. Machine learning has since taken the world by storm and I’ve been happy to use this skillset in projects for my clients!”
WHAT YOU’LL LEARN
HOW TO ADD MACHINE LEARNING TO YOUR WEBAPPS
Across the ten sections of this course, you will learn:
- Machine Learning and Neural Network concepts
- How to install and run TensorFlow.js in the web browser and Node.js environments.
- What is a Tensor? How to create them and run math operations on Tensors using the APIs
- How to manage memory in TensorFlow.js and why this is necessary.
- Data preparation for machine learning – We learn to assess a data structure to select input features and output labels, then we see how to import, normalise, shuffle, and split data. We also visualise data with a scatter plot.
- Defining a machine learning model using the TensorFlow.js Layers API. We see how to define layers, inspect, and compile the model.
- Training a machine learning model – how the structure of a model impacts training process including training speed, and accuracy of the model. We visualise the training process using the tfjs-vis library. We see how to evaluate the model using validation and testing data sets.
- Integrating TensorFlow.js into a user interface. Saving and loading models via the UI.
- How to make predictions with a trained model. We see how to visualise predictions using scatter plots and heatmaps.
- How to adapt a model to tackle a number of the major machine learning problem types – linear regression, binary classification, and multi-class classification.
LEARN-BY-DOING WITH REAL DATA
REAL-WORLD EXAMPLES USING HOUSE PRICES
Throughout the course we work with practical example projects, covering a range of machine learning problem types. All the examples are based on a house price data set from King County, USA – which includes Seattle! We use the data set to predict house prices and classify houses based on price and living space. In the course labs, you will work through the projects yourself, with additional challenges to deepen your learning.
We cover three projects, each covering a different type of machine learning problem:
- Linear Regression – We predict house prices from the living space of a house. As our first project, we explain all stages of machine learning process in detail, covering data preparation, machine learning models, training, validation and testing, and prediction.
- Binary Classification – We use house price and living space to predict which properties are on the waterfront! At this stage we are working with more complex models with multiple input features and a multi-layer neural network structure.
- Multi-class Classification – We use house prices and living space to predict the number of bedrooms in a house. We explore a number of techniques used in more complex machine learning models.
START LEARNING TODAY
UPGRADE YOUR SKILLS FOR THE MACHINE LEARNING REVOLUTION
Frequently Asked Questions
How long is the course?
The course is split into 10 sections, each with about 45mins video.
To get the most out of the course you should complete the labs and quizzes, and go back over any videos that you are unsure of. In total we estimate 50 hours to complete the course including all the labs – but it could be longer or shorter depending on your prior programming experience.
What skills do I need to take this course?
This is not a math-heavy course, so we don’t require any advanced mathematical knowledge.
You do not need any prior experience with machine learning in order to take this course. For those who do have machine learning experience, we are confident you will find the course stimulating. Wherever possible we teach machine learning concepts in a practical context using examples in TensorFlow.js.
Do I need any special equipment or software to take this course?
Any Mac, Windows, or Linux PC which can run modern web browsers can be used for this course. For some of the machine learning labs, the training process takes a long time to complete.
All the software we use in the course is available for download free of charge. In the videos we use the Chrome web browser and the Atom text editor, but you can use any modern browser and any text editor or IDE.