{!assets/text/instructor_info.md!}
In this course, students will study the emerging field of tiny machine learning (tinyML). This field is at the intersection of machine learning (ML) applications and embedded devices/microcontrollers. It requires both software and embedded-hardware knowledge. More specifically, students will follow a hands-on learning approach with training and optimizing ML models in such ways that they are deployable onto tiny microcontrollers. The course will involve work with an Nano 33BLE sense microcontroller.
There is no textbook requirement for this class. However, students are required to purchase an Arduino Tiny Machine Learning Kit. The cost of the kit is $60. If you get this from Amazon, you might be able to get it via free shipping.
Assessment | % of Final Grade | CSLO | MSPO |
---|---|---|---|
Quizzes | 30% | 1,2 | 1,2 |
Labs | 30% | 1,2,3,4 | 1,2 |
Project | 30% | 1,2,3,4 | 1,2 |
Class Participation | 10% | 1,2,3,4 | 1 |
D grades are not used. Refer to the Graduate Catalog for description of NG (No Grade), W, & other grades.
Labs/Project milestones that are late are assessed a 10% per day late penalty. Saturday and Sunday are each days.
{!assets/text/policy.md!}
Week | Topic | Assessments |
---|---|---|
1 | Introduction | - |
Machine Learning Paradigm | - | |
Building Blocks of Deep Learning | - | |
Exploring Machine Learning Scenarios | - | |
Building a Computer Vision Model | - | |
Responsible AI | - | |
AI Lifecyle and ML Workflow | - | |
2 | ML on Edge: Tensorflow Lite and Quantization | - |
ML on Edge: Post Training Quatization | - | |
ML on Edge: Quantization Aware Training | - | |
ML on Edge: Model Conversion and Deployment | - | |
Keyword Spotting | - | |
Visual Wake Words | - | |
Anomaly Detection | - | |
3 | Data Engineering | - |
Setting up hardware | - | |
Embedded hardware and software | - | |
Tensorflow Lite Micro | - | |
Keyword Spotting and Dataset Engineering | - | |
Visual Wake Words/Person Detection | - | |
4 | Gesture Recognition | - |
Introduction to Project | - | |
Responsible AI Deployment | - | |
DNN Compression | - | |
5 | Scaling TinyML | - |
History of TinyML | - | |
ML Hardware Acceleration | - |