Sensors Midwest 2018: Machine Learning Needs To Shake, Rattle & Roll

It’s a well-known fact that machine learning is the milk and money of artificial intelligence. Since artificial intelligence is going to play a significant role in just about every technology, it stands to reason that machine learning will be along for more than just the ride.


Industrial applications are essentially on the forefront of being automated with artificial intelligence at the helm in most. Obviously, industrial apps are noted for being rugged, involving motion, high torque and force levels, and a whole lot of vibration and shock. It has been found, by reliable resources, that, despite achieving excellent fault classification performance, recent deep-learning based models fail to perform well on vibration data from shaft speeds that they were not exposed to during training. In other words, machine learning is not quite picking up on the good, and bad, vibrations just yet.

Free Monthly Newsletter

Compelling read? Subscribe to FierceEmbeddedTech!

The embedded tech sector runs the market’s trends. FierceEmbeddedTech subscribers rely on our suite of newsletters as their must-read source for the latest news, developments and analysis impacting their world. Sign up today to get news and updates delivered to your inbox and read on the go.


As always, there are solutions, you just have to know where to find them. On Wednesday, October 17, 2018 from 10:10 am to 11:00 am, at Sensors Midwest in Rosemont, IL, machine learning engineer Erfan Azad at Uptime Solutions and software design manager Kim Luke will present a session titled, “Deep Learning Techniques for Vibration Fault Detection.” Their session addresses the aforementioned issues by various transfer-learning techniques to increase the robustness of the fault classifiers when encountering vibrations from shaft speeds that vary significantly from the ones used during their training.


Erfan earned a Bachelors Degree in Computer Science from Colby College and a Bachelors Degree in Engineering from Thayer School of Engineering at Dartmouth College. His research spans over areas of Control Theory, System Identification, and Machine-Learning. He is currently holding a position as a Machine Learning Engineer at Uptime-Solutions where he is working on fault detection and condition monitoring of bearings using machine-learning and deep-learning tools.   


Kim Luke is the design manager and director of data science for Uptime Solutions, a company focused on developing wireless products for the diagnosis and monitoring of machinery health. Kim has several years’ experience in software development, signal processing, and analyses in both academic and industrial settings. She earned both a Bachelors Degree in Mathematics and Masters Degree in Electrical Engineering from the University of Wyoming.


If industrial applications involving vibration analysis are your bread and butter, then there’s no reason or excuse for not attending Sensors Midwest and the session “Deep Learning Techniques for Vibration Fault Detection” as well as the other sessions in the Machine Learning & AI track. So, what are you waiting for? Register today.

Suggested Articles

Critics are concerned about a false sense of public health safety when temperature scanning is used in hospitals and other settings

Machine learning challenge will look for vocal communication between elephants and other behaviors

Iowa State University researchers are working with NSF grant