Production Readiness of Machine Learning for Industrial AI

General artificial intelligence (AI) research focuses on developing new machine learning algorithms and models with human-like cognitive capabilities. Industrial AI is more concerned with the application of ML models to address industrial pain-points for customer value creation, productivity improvement, and insight discovery. Industrial AI proposes a promising framework for future industrial control systems, but manufacturers and solution partners still need to understand how to implement and integrate an AI model into the existing industrial control system. A well-trained ML model provides a number of benefits and opportunities for industrial control optimization; however, a "terrible" industrial AI design and integration also limits the capability of an ML model. In this talk, I'll present the top challenges (Domain knowledge transfer, data preprocessing, ML model accuracy, coexistence of traditional control and AI control, real-time constraints) that manufacturers and their AI partners will face when implementing an AI solution. I will then provide real-world examples of successes and will explain how to address the obstacles for a successful AI deployment.
Who is this presentation for?
Industrial manufacturers and AI solution partners looking to deploy industrial AI solutions.
Prerequisite knowledge:
A basic understanding of AI and ML.
What you'll learn?
Key Takeaways: 1. The high value of applying AI in manufacturing plants. 2. How industrial AI is different from traditional AI research thus needing a different approach. 3. Recommendations on how to effectively deploy AI in manufacturing plants.

Dr. Hongbo Zou is a staff engineer and technical leader at Petuum Inc. Before joining the Petuum team, he worked on the project of SDN datapath validation in VMware Inc. for over four years. Additionally, he worked in high-performance computing research for several years at DOE national labs. His primary areas of focus include machine learning, large-scale data analysis, and high-performance computing. Currently, he is working on a research project that studies how to validate and enable machine learning in industrial AI. Hongbo Zou received his Ph.D. from Queensland University of Technology and his M.S. from Georgia Institute of Technology.