Overview
This course covers the learning outcomes and goals of understanding recent advances in machine learning as applied to the Internet of Things (IoT). It aims to teach the differences between physics-based and data-driven approaches to IoT, challenges in applying data-driven approaches, and how machine learning can address these challenges. The course also focuses on failure prediction, prognostics, and real-case studies illustrating the application of deep learning, gradient boosting, transfer learning, and other machine learning techniques for IoT applications. The teaching method includes theoretical overviews, discussions on challenges and advancements, and a real-case study. The intended audience for this course includes individuals interested in machine learning, IoT, industrial equipment, and data analysis in the context of sensor technologies.
Syllabus
Intro
Overview of loT
Buckets of Use Cases
Use case: fleet management
Use case: on-time- in-fill shipping
Failure prediction example
What is the value of prognostics?
Challenge Connectivity
Failure and Repair Data Unreliable and
Failures Dispersed Across Many Types
Large Quantity of Signal Data
Challenge Sensor Limitations
Strengths
Neural networks
Transfer learning
Taught by
Open Data Science