Introduction to Edge Machine Learning
Machine learning is on the edge of revolutionizing our lives by automating and simplifying work processes and improving several industries to our benefit. For instance, in automotive sector machine learning will have the biggest impact on identification and navigation of roads and obstacles in real-time for autonomous driving. We will see the more heavy use of machine learning running on edge embedded systems also in robots in manufacturing facilities.
This course will boost your know-how towards a concrete use of machine learning by giving an introduction what machine learning is all about and especially its role in embedded edge devices. During the course we retrain an image recognition model that detects objects. This training course includes hands on exercises and we will use Google’s Tensorflow and Keras in a container on your own laptop and a Raspberry Pi with a camera attached.
During the training course we will use Python. Prior Python experience is not necessary but basic programming concepts understanding is required. You should also have minimal prior experience using Linux/UNIX as a user.
PRACTICAL EXERCISES / TOOLS
Approximately half of the time will be on hands-on exercises. They have been designed to highlight the development process of machine learning.
We use a PC as a host. On the target we run Linux.
You will keep the Raspberry Pi 3, camera and case, to be able to continue exploring machine learning after the training.
- What is machine learning?
- Material and tools used during training.
- Basic terminology walk through
- Process for training and executing
- Industry and ecosystem overview
BASIC MACHINE LEARNING
- Training a model
- Test and validation in machine learning
- Basic tuning of a model
- Linear Algebra basics
- Cost Function
- Gradient Descent
- What is deep learning
- Introduction to various models
- Model chaining
INFERENCE PERFORMANCE ON EDGE DEVICES
- Reducing accuracy of model in exchange of performance
- Performance comparison of different computing hardware
DATASCIENCE UNDERSTANDING NEEDED FOR MACHINE LEARNING
- Extracting knowledge and insights
- Training with sequential data
- “Debugging” a model using Tensorboard
FURTHER LEARNING WHAT TO DO AFTER THIS COURSE
- Intermediate terminology and concept walk through to aid in further learning in the field