Keynote I

With so much data available from connected products, how do we do something impactful with it? Everyone's go-to answer, Deep learning? There will be cases where neural nets are the best approach - but if you can identify the problem, better to use deterministic math and physics instead of a black box to identify a failure precursor and take action.

In this talk, we will touch on numerous relevant practices for improving field product quality and reliability:

• Consuming real time telemetry and feeding into an on-board damage model, such as Tire wear indicator.
• Leveraging connected product to self-test as it's being manufactured
• Using large data sets and computer vision to identify quality defects before it gets out of the factory
• Safety beyond the stars - using fleet data to improve regulations and standards

Hope this presentation will be beneficial to both academia researchers as well as industrial professionals as a look at practical engineering application of Connected Devices, Big Data, and Predictive Health Monitoring.

Wes Morrill is currently the Director of Reliability, Test and Analysis at Tesla, Inc.

Since joining in 2011, he has over 12 years of technical and management experience at Tesla covering various aspects of product development including component and system physical testing, design for reliability, and computer-aided engineering analysis. He has worked on every product in the portfolio (in approximate order): Roadster, Mercedes Smart Car, Mercedes B-Class, Toyota Rav4 EV, Model S, Model X, Model 3, Powerwall, Supercharger, Model Y, Megapack, Semi Truck, Solar Roof, and is currently the lead engineer for Cybertruck.

Born and raised in Danbury, Connecticut, Wes graduated in 2008 from Carnegie Mellon University, majoring in Mechanical Engineering, minoring in robotics and industrial design.

Keynote II

Cybercrime now costs trillions of dollars annually. Machine learning can help to detect cyberattacks in big data, but it needs to overcome several challenges to be effective. Like the proverbial “needle in a haystack” analogy, severe class imbalance can cause machine learning classifiers difficulty. To compound this problem further, class rarity is not uncommon where only very few instances are available from the positive class causing classifiers further trouble in being able to discriminate between the classes. We evaluate applying sampling techniques to treat the class imbalance problems for detecting cyberattacks in big data. Properly preprocessing the input data is an important step, and we discuss data quality issues of the input features. Finally, we cover how feature selection can also be helpful for detecting cyberattacks in big data.

Dr. Taghi M. Khoshgoftaar is Motorola Endowed Chair Professor of the Department of Computer and Electrical Engineering and Computer Science, Florida Atlantic University and the Director of NSF Big Data Training and Research Laboratory.

His research interests are in big data analytics, data mining and machine learning, health informatics and bioinformatics, social network mining, security analytics, fraud detection, and software engineering. He has published more than 850 refereed journal and conference papers in these areas. He was the conference chair of the IEEE International Conference on Machine Learning and Applications (ICMLA 2016 and ICMLA 2019). He is the Co-Editor-in Chief of the journal of Big Data. He has served on organizing and technical program committees of various international conferences, symposia, and workshops. Also, he has served as North American Editor of the Software Quality Journal and was on the editorial boards of the journals Multimedia Tools and Applications, Knowledge and Information Systems, and Empirical Software Engineering and is on the editorial boards of the journals Software Quality, Software Engineering and Knowledge Engineering, and Social Network Analysis and Mining.

For his selected publications, please see his Google Scholar link: https://scholar.google.com/citations?user=-PgNSCAAAAAJ&hl=en