AI on the PC

Summary

Learn how to use Intel® hardware, software, and solutions for AI on the PC. Solve the difficulties of deep learning inference on edge devices.

By the end of this course, students will have practical knowledge of:

  • Windows* Machine Learning to accelerate machine learning applications
  • The Model Optimizer and inference engine in the Intel® Distribution for OpenVINO™ toolkit on multiple types of hardware
  • Deep learning tools and frameworks, such as TensorFlow* and Open Neural Network Exchange (ONNX*)

The course is structured around eight modules of lectures and exercises. Each module requires one hour to complete.

Prerequisites

Python* programming

Calculus

Linear algebra

Basic statistics

Module 1

This class introduces the basics of AI:

  • Applications of AI and ways it can transform industries
  • Comparison between machine learning and deep learning
  • Basic deep learning terminology
Module 2

This class reviews how Intel hardware is used for AI. Topics include:

  • Intel's vision for AI on PC hardware and software
  • How different hardware addresses various AI tasks, such as training and inference
  • The analytics ecosystem, which is made up of toolkits, libraries, solutions, and hardware
Module 3

This class teaches about deep learning frameworks and provides:

  • An overview of the optimized frameworks for machine learning
  • An introduction to TensorFlow and central concepts, such as computational graphs and sessions
  • Instructions to create and run a simple computational graph in Python
Module 4

This class explains the end-to-end AI training workflow. Topics include:

  • How to clean, normalize, and optimize a dataset
  • An example of how to train a GoogLeNet Inception neural network model 
  • How to evaluate a trained model and test it for accuracy and performance
Module 5

This class introduces the challenges of AI inference at the edge. Topics include:

  • What edge computing is and how it will influence modern technology
  • The importance of inference on the edge and why it's required by emerging markets
Module 6

This class introduces how to use Windows Machine Learning to accelerate AI development. Topics include:

  • The benefits of using Windows Machine Learning for inference on the edge
  • How to improve performance using the most popular frameworks with ONNX models
  • How the Windows Machine Learning stack can improve performance of AI models on integrated graphics
Module 7

This class introduces the Intel Distribution of OpenVINO toolkit and how to use it to run inference on the edge. Learn about:

  • The different parts and advantages of using the toolkit
  • How to use the Model Optimizer to improve the model topology of pretrained networks 
  • How to use the inference engine to run on different types of hardware
Module 8

Complete this course with a review of the previous topics, including:

  • How Intel hardware, toolkits, and solutions allow developers to create applications for AI on the PC
  • Why Intel's collaboration with Microsoft* improves deep learning performance for PCs through Windows Machine Learning
  • An introduction to Intel Distribution of OpenVINO toolkit to use with deep learning frameworks for powerful AI applications