Explore the convergence of Machine Learning and IoT in this comprehensive conference talk. Delve into the concept of "Edge ML" and its potential to revolutionize IoT security and efficiency. Learn about the challenges of cloud-based ML in IoT applications and discover how modern embedded systems and microcontrollers are enabling on-device machine learning. Gain practical insights into implementing Edge ML in IoT projects, including examples using Google Coral, Raspberry Pi, and microcontrollers. Understand the benefits of bringing machine learning closer to sensors and IoT devices, such as improved security, faster insights, and predictive capabilities. Examine the state-of-the-art in Edge ML and IoT, covering topics like TensorFlow Lite, federated learning, and the use of single-board computers for machine learning tasks.
Overview
Syllabus
Intro
Zeke
Washing Machine
ML vs Deep Learning
The Streetlight Problem
Human vs Machine Learning
Discipline of Machine Learning
Machine Learning Tools
Example
Train predict tune loop
Train predict model
Train predict classifier
Deployment
Pretrained APIs
Azure Face API
Cloud vs Local Performance
The IoT is the source of data
Volume
The problem
The solution
What we want
What does this look like
What you get instead
Google Coral
Google Coral Demo
IoT
Raspberry Pie
Lowcost ML
Microcontrollers
Federated Learning
Three Big Things
Single Board Computers
Tensorflow Light
Taught by
NDC Conferences
