THREE TAKEAWAYS FROM THE INDUSTRIAL AI FOR SMART MANUFACTURING VIRTUAL CONFERENCE
SEPTEMBER 17, 2020
Predictronics had a productive experience participating in the Industrial AI for Smart Manufacturing Virtual Conference.
Hosted by our Predictronics co-founder, Dr. Jay Lee, and the Foxconn iAI Institute, this virtual conference brought together panelists from across industry and academia to discuss the impact of Industrial AI on smart manufacturing, the power of digitalization within the industrial world, and the new era of manufacturing in a post-pandemic world.
CTO Dr. David Siegel was a keynote speaker during the conference, presenting successful Predictronics case studies and demonstrating the real-world application and value of AI-enabled, data-driven predictive solutions within manufacturing.
Dr. Siegel shares his top three takeaways from the virtual conference:
1. Recent Advancements in Humble AI
One exciting new development discussed was the concept of humble AI. As AI-based solutions are deployed for various industrial applications, the trust in the AI models becomes more important. Humble AI helps determine whether the trained model is outside the region it was trained with and what action should be taken if this occurs. For example, does it suggest no recommendation, or does it default to an alternative or simpler model in that case? Humble AI seeks to bring greater understanding in this area.
The topic of uncertainty has been popular in recent years when developing algorithms and analyzing results from AI/machine learning models. To be most effective, an algorithm should provide honest feedback on how confident it feels about the end results of its data insights. Ultimately, this will help influence what decisions or recommendations are presented to the user of the software.
2. Dissemination of Knowledge in Industrial AI Training
Two other important topics examined by panelists include the dissemination of prognostics and health management (PHM) and Industrial AI (IAI) technology to the larger community and the resources necessary to learn the fundamentals in the field. Representatives from the PHM Society emphasized education as part of their core mission, demonstrating the various tutorials, training courses, panel sessions, and data challenge competitions they offer to help facilitate further learning in regard to PHM topics. In addition, Foxconn highlighted their new Foxconn iAI Institute, which will provide training courses, data challenge competitions, learning modules, and past data sets on Industrial AI.
By increasing training efforts, various organizations are now providing a better understanding of the limitations and feasibility of PHM and IAI technologies to others. This education will hopefully address current issues industry professionals face when adopting and implementing these technologies.
3. Importance of Real-World Success Stories in Industrial AI
What seemed to resonate most with attendees were the demonstrations of real-world AI solutions and successful deployments for industrial applications. Various successful case studies on industrial robotics, stamping presses, high pressure die casting, electronics manufacturing, aviation, and oil/gas were presented. These cases showcased the great advancements and maturation of AI technologies in recent years and examined the methodologies and strategies for implementation.
One challenging aspect many solutions providers face currently is quantifying the success of these technologies. This can prove difficult as getting clear ROI numbers from end-users is not always straightforward with many companies leery of divulging the value and benefits for fear of showing their hand to competitors. Despite this challenge, panelists were able to present a few examples highlighting improvements in OEE (5% - 10%) and, in some cases, millions of dollars saved from the Implementation of these solutions.
Other noteworthy speakers included representatives from the World Economic Forum, NIST, GE, PARC, the University of Cincinnati, and the PHM Society. Predictronics is proud to have been a part of the conversation with such distinguished colleagues and intellectuals, fostering a dialogue on the importance of Industrial AI in the future of manufacturing.
For more information and weekly updates, follow Predictronics:
Know What Happens Next with PDX!
TRY IT TODAY