Applied Fundamentals of Industrial AI, Machine Learning and IIoT by Rajiv Anand, Quartic.AI, Xiaozhou Wang, Quartic.ai

Intended audience:
Maintenance and Reliability engineers, managers and leaders; industrial process control, automation, MES and operational excellence professionals.

The workshop will serve as a primer for those who have little to no knowledge of data science, AI, and IIoT; and provide a concise, structured learning and reinforcement of topics for those with some basic understanding and knowledge.

PART I – FUNDAMENTALS

The basics

  • Data vs. insights – how do we extract insights from data
  • What is artificial intelligence; the type of problems it is solving.
  • What is machine learning and data science, how does it work and how is it different from traditional logic and rule based systems
  • Learning methods: How does an algorithm learn:
    • Supervised Learning
    • Unsupervised Learning
    • Reinforcement learning

    Algorithms

  • Different algorithm types and their application:
    • Regression
    • Classification
    • Clustering
    • Gausian models
    • Neural networks and deep learning
  • Patterns in data: pattern recognition, anomaly detection
  • Machine Learning implementation process

  • Data types, basic exploration of data with statistical methods
  • Co-relation and causation
  • What are features, feature selection, extraction and importance
  • Algorithm selection
  • Model training, validation and optimization
  • Model performance evaluation and tuning
  • Feedback mechanisms
  • Quiz: Fundamentals reinforcement

    PART II: IMPLEMENTATION

    Deployment of Models:

  • Data processing basics – batch and streaming data processing
  • Cloud, Fog and edge
  • Data pipeline and IIoT architectures
  • Security
  • Hands-on demonstration and exercise:

  • Using a modern industrial IIoT system, a step-by-step demonstration of building machine learning applications for anomaly detection and health prediction by an industrial user with no data science or programming experience.
  • Model Building Exercise:

  • Limited number of users will be able to build machine learning models for a prediction of faults and performance of a hydraulics system using a problem set and data set provided for the workshop