Nowadays, you may have heard of many keywords like Embedded AI /Embedded ML /Edge AI, the meaning behind them is the same, I.e. To make an AI algorithm or model run on embedded devices. Due to a massive gap between both technologies, techies don't know where to start with it.
Getting Started with Embedded AI | Edge AI
3h 33m 42s
English
Paid
So we thought to share our engineer's experience with you via this course. We have created an application to recognize the fault of a motor based on the vibration pattern. An Edge AI node developed to perform the analysis on the data captured from the accelerometer sensor to recognize the fault.
We have created detailed videos with animation to give our students an engaging experience while learning this stunning technology. We assure you will love this course after getting this hands-on experience.
The Motivation behind this course
One and half years back, It was surprising when techies heard of the embedded systems running standalone Deep learning model. We thought to design an application using this concept and share the same with you via this platform.
How to start the course?
There are two possible ways to start this course. We have divided this course into Conceptual Learning and Practical Learning. You can either jump directly to the Practical videos to keep the motivation to learn and later can go to fundamental concepts. Or you can start with the basic concepts first then can start building the application.
What you will get after enrolling in the course
1. You will get Conceptual + Practical clarity on Embedded AI
2. After this course you will be able to build similar kind of applications in Embedded AI
3. You will get all the Python scripts and C code(stm32) for Data capturing ,Data Labeling and Inference.
4.You will be able to know in depth working behind the neural networks
Requirements:
- Knowledge of C or Python Language is plus
- Knowledge of stm32 is plus
Who this course is for:
- Embedded AI Explorer
- Embedded Enthusiast
- Engineers
- Artificial Intelligence/Deep learning Enthusiast
- M-Tech/PhD Students
What you'll learn:
- Learn basic concept behind AI/DL
- Learn how to use KERAS deep learning library in python?
- Learn how to capture and label data from sensors via Microcontroller
- Learn to create a Neural network and how to train them on data
- Learn to implement Deep learning model on a microcontroller and can run inference on it.
About the Author: Udemy
Udemy is the largest open marketplace for online courses on the internet. Founded in 2010 by Eren Bali, Oktay Caglar, and Gagan Biyani and headquartered in San Francisco, the company went public on the Nasdaq in 2021 under the ticker UDMY. The platform hosts well over two hundred thousand courses across software development, IT and cloud, data science, design, business, marketing, and creative skills, taught by tens of thousands of independent instructors. Roughly seventy million learners use it worldwide, and the corporate arm — Udemy Business — supplies a curated subset of that catalog to enterprise customers.
Because Udemy is a marketplace rather than a single editorial publisher, the catalog is uneven by design. The strongest material lives in the long-form, project-based courses authored by working engineers — full-stack JavaScript, React, Node.js, Python data science, AWS, Docker and Kubernetes, mobile development with Flutter and React Native, and cloud certification preparation. The CourseFlix listing under this source is the slice of that catalog that has been mirrored here for offline-friendly viewing, organized by topic and updated as new releases land. Pricing on Udemy itself swings dramatically with the site's near-permanent sales, which is why the platform is best treated as a deep reference catalog: pick instructors with strong reviews and a track record of updating their material rather than buying on the headline price alone.
Watch Online 51 lessons
0:00
/ #1: What is an Artificial intelligence?
All Course Lessons (51)
| # | Lesson Title | Duration | Access |
|---|---|---|---|
| 1 | What is an Artificial intelligence? Demo | 04:36 | |
| 2 | What is Machine Learning? | 02:09 | |
| 3 | What is Deep Learning? | 03:53 | |
| 4 | What is an Embedded/Edge AI? | 04:53 | |
| 5 | Applications of Embedded AI | 02:54 | |
| 6 | Overview of the Tools used. | 01:47 | |
| 7 | What is Tensorflow? | 06:11 | |
| 8 | What is Keras? | 03:27 | |
| 9 | Comparison between Keras and Tensorflow | 05:33 | |
| 10 | Installation of Keras and Tensorflow | 01:22 | |
| 11 | What is STM32 and X-CUBE AI | 01:55 | |
| 12 | Development Board used | 01:14 | |
| 13 | What is Supervised Learning? | 02:13 | |
| 14 | What is Unsupervised Learning? | 01:59 | |
| 15 | Artificial Neuron Vs Real Neuron | 02:19 | |
| 16 | What is an Artificial Neural Network? | 02:36 | |
| 17 | What are layers and Forward propagation in NN | 04:30 | |
| 18 | What is an Activation Function? | 03:57 | |
| 19 | What is Gradient and Gradient Descent? | 03:40 | |
| 20 | Optimization Algorithm and Loss function | 04:24 | |
| 21 | How a Neural Network Learns? | 04:27 | |
| 22 | The Concept of Loss functions in detail | 02:56 | |
| 23 | The process of training and testing a NN | 05:00 | |
| 24 | Why Overfitting occurs in NN and How to avoid it? | 04:45 | |
| 25 | Why Underfitting occurs in NN and How to avoid it? | 03:29 | |
| 26 | Hyperparameter of NN -> Learning Rate | 03:16 | |
| 27 | What is Batch and Batch size of a Training samples? | 03:19 | |
| 28 | Transfer Learning and Fine tuning Hyperparametrs in NN | 05:21 | |
| 29 | What is Convolution? | 06:06 | |
| 30 | What is a Convolution Layer in NN? | 04:42 | |
| 31 | What is Max Pooling Layer? | 03:58 | |
| 32 | What is Dropout layer? | 01:44 | |
| 33 | One Hot Encoding of Output Classes or Labels | 06:07 | |
| 34 | What is Confusion Matrix? | 03:53 | |
| 35 | Difference between with or without normalization Confusion matrix | 01:57 | |
| 36 | Introduction To Python and Writing first Program | 06:25 | |
| 37 | Inroduction to Numpy Package | 05:23 | |
| 38 | Introduction to Pandas Package | 04:20 | |
| 39 | Introduction to Matplotlib | 02:00 | |
| 40 | Key Steps for the implementation of Edge AI | 03:27 | |
| 41 | Accelerometer Sensor Module | 02:34 | |
| 42 | C code to capture data from Accelerometer | 14:28 | |
| 43 | Python Script to Collect and Save Data in Binary file | 08:52 | |
| 44 | Python script to Clean and Label Data | 05:53 | |
| 45 | Defining a Convolution Neural Network to Learn from Captured Data | 05:10 | |
| 46 | Python Script to Train the Neural Network | 11:08 | |
| 47 | How we captured data and trained the model on it | 02:10 | |
| 48 | Performance Evaluation of the Model (Plotting Confusion Matrix) | 02:22 | |
| 49 | Convert KERAS model to c code | 06:54 | |
| 50 | Integration of generated c code to acccelerometer module code | 02:53 | |
| 51 | Infer the Fault State on the machine (demo) | 03:11 |
Unlock unlimited learning
Get instant access to all 50 lessons in this course, plus thousands of other premium courses. One subscription, unlimited knowledge.
Learn more about subscriptionRelated courses
-
Updated 11mo agoPython for Data Engineers
By: Andreas KretzIf you want to take your skills in Data Engineering to the next level, you are in the right place.2h 21m5/5 -
Updated 1y agoDeep Learning: Advanced Computer Vision
By: UdemyThis is one of the most exciting courses I’ve done and it really shows how fast and how far deep learning has come over the years. When I first started my deep15h 10m -
Updated 11mo agoDockerized ETL With AWS, TDengine & Grafana
By: Andreas KretzData engineers often need to quickly set up a simple ETL script that just gets the job done.29m
Frequently asked questions
What prerequisites should I have before taking this course?
Before enrolling, you should have a basic understanding of artificial intelligence and machine learning concepts. Familiarity with programming, especially in Python, is recommended since the course involves writing Python scripts and handling packages like Numpy, Pandas, and Matplotlib.
What will I build during the course?
The course involves building a project that captures data from an accelerometer sensor, processes it using Python scripts, and trains a Convolutional Neural Network to analyze the data. The trained model is then converted to C code and integrated with the accelerometer module code to infer the fault state on a machine.
Who is the target audience for this course?
This course is suitable for technology enthusiasts, developers, and engineers interested in learning about embedded AI and its applications. It is particularly beneficial for those who want to understand how to implement AI models on embedded devices like the STM32 development board.
What specific tools or platforms will I learn about?
The course covers tools such as TensorFlow and Keras for building AI models. You will also learn about the STM32 development board and X-CUBE AI, which are used to deploy AI algorithms on embedded devices.
What topics are not covered in this course?
The course does not cover advanced topics like reinforcement learning or more complex neural network architectures beyond convolutional neural networks. It focuses on the basics required for running AI models on embedded systems.
How much time commitment is required to complete the course?
The course consists of 51 lessons. Although the total runtime is not specified, you should expect to invest several hours to go through the material, complete exercises, and work on the project, depending on your familiarity with the topics.
How does this course benefit my career in AI or embedded systems?
By completing this course, you gain practical skills in implementing AI models on embedded devices, a growing area of interest in the tech industry. This knowledge can be valuable for roles that involve developing smart IoT devices or edge AI solutions, providing a foundation to explore more advanced AI and machine learning courses.