by Dr. Gerhard Luhn, on November 20, 2018
- Industry 4.0
- Digital Transformation
- Smart Manufacturing
by Dr. Gerhard Luhn, on November 20, 2018
At the XXX Advanced Process Conference in Austin, Texas, microelectronics specialists set to work with artificial intelligence. Considering the enormous innovations in the field of microelectronics in the last 30 years, no AI system can be credited with comparable innovative power on its own yet with regard to ‘autonomous creativity’. Instead, microelectronic specialists are now focusing on methodologically better integrating their ‘subject matter expertise’ into the further roadmap of microelectronics development - also using AI systems. However, new information processing concepts are emerging alongside AI, and some of those are developed by SYSTEMA.
The XXX. 2018 Advanced Process Conference in Austin, Texas, recently concluded. For more than 30 years, this conference has assembled microelectronics specialists from around the world to report on interesting, sometimes spectacular, innovations and discuss what we’re likely to see in the future. The conference has been organized by James Moyne for many years.
It was not until May of this year that Google proclaimed that the entire focus of its research activities would be on AI. The federal government of Germany also published a cornerstone paper outlining key points of their ‘new AI strategy’ in July 2018.
Against this background, this APC conference was remarkable. After all, one has to think that those designers and producers who invent and produce exactly the substance and ingredients that make up the dreams of AI-tinkerers themselves are deeply intertwined in the AI web. Deep learning was one element of AI in the spotlight of this conference.
However, the main topic of the working session of the conference was: ‘Subject Matter Expertise’ and its importance and role in relation to AI. As a result, the role and promise of AI is seen with some skepticism. In particular – and this was discussed in detail with direct examples from the field of microelectronics – for the chipmaker specialists who indicated that AI seemed to lack any innovative power or imagination. The experts in microelectronics have proven their worth for decades with some groundbreaking innovations (supported by James Moyne in the IRDS / International Roadmap for Devices and Systems). Specifically, this innovative power and imagination enable complex systems which are necessary for truly deep learning. The failure of ambitious projects – such as the failed use of the IBM AI system ‘Watson’ in the German Cancer Research Center Heidelberg – documents this problem.
With this in mind, Gerhard Luhn and Gerald Hüther contributed to this conference and with the discussion of a new, physically and logically inspired approach. They propose (Luhn/Hüther 2017¹; Luhn/Hüther 2018 [forthcoming]²):
This blindness to imaginary states of life is not only homegrown by AI research, but seems to reflect a basic attitude of many sciences. After all, every natural law is, so to speak, "not determined by this world" but is determinate in the literal sense of the word forever (lat. determinare, determinate).
In reality – as suggested by the many-worlds interpretation of quantum mechanics – people and especially curious engineers exploit fundamental primary activity, which is not covered by classical physics³. As a result, engineers may discover new phenomena and build phenomenal models in order to simplify and, nevertheless, successfully drive complex processes and products.
As an example, SYSTEMA’s approach for user-centric automated failure recognition (USE) reflects this philosophy. Here, USE provides a framework which allows the integration of new algorithms and classification methodologies of growing complexity in a short time frame.
Another example is SYSTEMA’S research activity entitled ‘Real-time information system RI-Suite’. The underlying methodology supports a consistent and intuitive mapping of the user’s production world into a mathematical model, which enables system performance similar to quantum computing. The aim of this approach is to bridge the gap between the growing amount of production data and analyzing this very data in real-time. Production sites are continuously pushed to optimize productivity, and RI-Suite offers standard methods of fab optimization, as well as previously unreachable functionalities like continuous production forecasting.
The good news is that microelectronics specialists have access to, or continually generate, exactly this knowledge that is exemplary of their open attitude and participation. Unfortunately, this structure is not reflected in the promises of an AI that has not yet discovered the importance of imaginary system states.
Consequently, the specialists from the APC workshop have included this subject matter expertise issue on their agenda for the next few years in order to make their contribution to a necessary humane dimension in modern-day projects. We can only recommend this approach – from a logically clear, and in no way romantically transfigured perspective – to be incorporated in the field of AI. Anyway, colleagues from AI: you are cordially invited to join the movement!
For additional information regarding our approach to enabling big data and deep learning through automation solutions, check out SYSTEMA's guide to digital transformation.
1. Luhn, G. and Hüther, G. (2017) ‘Thinking, future and ‘non’-causality. On life and consciousness in the complex plane’, Int. J. Foresight and Innovation Policy, Vol. 12, Nos. 1/2/3, pp.5–36
2. Luhn, G. and Hüther, G. (2018, forthcoming) ‘The oscillating mind-body-system (OMB). On the deep nature of human knowledge’
3. Schiemann, G. (2009) ‘Werner Heisenberg’s Position on a Hypothetical Conception of Science’; in: Heidelberger, M.; Schiemann, G. (ed.): ‘The Significance of the Hypothetical in the Natural Sciences’, Berlin 2009
Never miss a blog post again!