Industry Technologies
(Part 2 — To be continued in Issue #9 - March 2022)
“Artificial intelligence will greatly change the economy in the coming years,” Professor Erik Brynjolfsson predicted recently in an interview with the German weekly magazine ‘Der Spiegel’. The Director of Digital Economy Lab at the Stanford Institute for Human-Centered continued: “However, this is less about the short-term ups and downs and more about fundamental changes. I am optimistic that we will see a boom in growth and productivity around the world.”
The human emotion of optimism thus as the best possible offer by humanoid intelligence of the premium level has for a prediction: a realistic assessment (vulgo: estimate). One of the many areas in which artificial intelligence (IT jargon: AI) though conceived by humans, is far superior to the natural intelligence by us humanoids.
But let's start at the beginning.
.jpg)
Erik Brynjolfsson. His research – these days at Stanford University, before at M.I.T., examines the effects of information technologies on business strategy, productivity and performance, digital commerce, and intangible assets.
Erik Brynjolfsson. His research – these days at Stanford University, before at M.I.T., examines the effects of information technologies on business strategy, productivity and performance, digital commerce, and intangible assets.
Doubtlessly we are only at the starting phase of a technological revolution. Learning machines will not only take over routine tasks but also decide on the loan for buying a house and make medical diagnoses. Learning machines enable the development of sustainability concepts regarding energy consumption in the housing segment – for our cities, optimized solutions for on-demand mobility (including semi- and fully autonomous vehicles) and logistics – to mention but a few applications one could think of.
AI-based algorithms can predict
May it be on the ideal candidate for a particular job or a political post. They will make decisions about us, for us, and with us – if we go about it the right way. Likewise, any fundamental invention AI bears major opportunities as well as indeed – risks. Therefore, we as global and local societies will have to respond comprehensively to the possible deployments in terms of labor, social, and educational policy. And yes, for all such applications that directly affect human beings, we will want to develop and implement control mechanisms. To give a typical example: The 'Compass' recidivism prediction software used by the U.S. justice system in the context of suspended prison sentences. (The data used here to feed the deposited algorithm is judged by various computer scientists to be extremely questionable.)
Strictly speaking, artificial intelligence is a bundle of mathematical methods for generating information from large amounts of data. In the literature, these methods are bundled into four to six classes. Thus, according to secondary literature, the following classes can be distinguished:1. Search and optimization2. Logic3. Probabilistic methods for uncertain reasoning4. Classifiers and statistical learning methods5. Artificial neural networks6. Evaluating progress
Garbage in – Garbage out
Data, data, and again data feed AI, so to speak. With their quantity and, above all, the quality of data, the efficiency of any algorithms to be generated stands and falls. Whereby algorithms are nothing more and nothing less than defined instructions for solving clearly defined mathematical problems.
Artificial Intelligence (AI) also machine intelligence (MI) is intelligence demonstrated by machines, in contrast to the natural intelligence (NI) displayed by humans and other animals. In computer science, AI research is defined as the study of ‘intelligent agents’.
As with any problem, an analysis of the influencing variables is first required. "Particularly with Big Data-based applications, the greatest care is required here," warns Prof. Dr.-Ing. Thomas Gries, who is the holder of the Chair of Textile Mechanical Engineering and Head of the ITA Institute for Textile Technology at the University of Excellence, RWTH Aachen in Germany.
.jpg)
Prof. Dr.-Ing. Thomas Gries
He leads a whole series of research projects for textile and other AI applications that are closely connected to various industrial applications. He states: “Among other aspects, the structure of the data must first be defined for an intended objective of each AI application – thus the quality.” And with a view to the risk situation, we all – this should also be said here – as private individuals as well as in the professional environment – certainly want to handle our personal as well as institutional data with more care.
What does AI offer?
Application areas that are emerging at present are the analysis and utilization of individual customer information and also qualitative/quantitative studies for future-oriented retail market research as well as in the investment goods sector. In the field of process automation for digital image processing and with the assistance systems (Industry 4.0 – Digital Manufacturing) neural networks are used on the testing level for determining setting parameters for machines and system installations. These methods are already being evaluated in research and industrial projects. Industrial use on a broad basis has yet to be established for the most part.
Zünd Cutter
As soon as it works, no one calls it AI anymore. This insight originates in the work of John McCarthy, one of the pioneers of artificial intelligence.
DemystificationBeyond predictive maintenance, as offered at the early-stage level by CAM (Computer-Aided Manufacturing) suppliers to sewn goods manufacturers with spreaders and cutters such as Bullmer, Lectra/Gerber, Pathfinder, Zünd, and others indeed AI intelligence has already become part of everybody’s daily life: from the navigation system in our cars to finding us the best, fastest or most beautiful route. AI takes into account the current road traffic to search engines: if, for example, Google does not provide the same answer for every user, but instead adapts results to the respective user profile. A best practice example to online shopping platforms making us aware ("Customers who bought, this item also bought ...").
AI – an apparel scenario3rd quarter 2021 global trade publication just-style.com headlined a feature rolling out results for dedicated research by Global Data Plc, today’s owner for the premium online business publication to the fashion industry. One of many aims for the study had been to display indications on where apparel is heading to leverage from what the technology can offer. The outcome from a general point of view: “The apparel industry is seeing an increase in artificial intelligence investment across several key metrics. AI is gaining an increasing presence across multiple sectors, with top companies completing more AI deals, hiring for more roles related to artificial intelligence”.
The London-based data analytics and consulting company’s thematic approach to sector activity seeks to group key company information on hiring, deals, patents, and more by topic to see which companies are best placed to weather the disruptions coming to their industry segment.
Findings at-a-glance
Hiring patterns within the apparel sector as a whole are pointing towards an increase in the level of attention being shown to roles related to the development and/or application of artificial intelligence. “There was a monthly average of 7,435 actively advertised-for open AI-related roles within the industry in April this year [remark by NEEDLE’S EYE editor: 2021], up from a monthly average of 5,703 in December 2020”, states the summary of the GlobalData methodical analysis. As a company-related result of the “Just Style" feature – recommended for detailed studying – the following excerpts may be quoted here: “According to GlobalData methodical analyses, LVMH, Nike, Foot Locker are classed as dominant players in AI in the sector, with an additional eight companies classified as leaders.”
From an inter-sectoral perspective, the clear edge in Artificial Intelligence seems to be kept by the U.S. since today’s largest talent pools are evidenced by a survey by LinkedIn from 2018. India accordingly occupies rank #2, with the worldwide second-highest number of AI professionals. Talented workforce playing a key role in technology advancement Europe as a whole certainly facing the challenge to catch up promptly.
Fashion business: Early adoptersAlthough still far from market penetration, the range of the first AI applications extends from buyer behavior and intelligent forecasting for demand on fabric/trims towards truly predictive machine/system maintenance.
Asos and Bohoo, Calvin Klein, H&M, and Jelmoli do it. The North Face, Patagonia, and Puma are on board. Tommy Hilfiger and Zara have jumped on the bandwagon of fashion/retail brands receiving profound support to solve sizing issues.
Collaborating with start-up types of companies like ZyseMe out of Berlin, Germany, a realistic narrowing down of the standard sizes with the expected best fit result becomes possible: It takes only 5-12 questions via their platform to figure out the look she or he adores in the size they need or prefer – as judgment to a good portion follows subjective criteria. No measuring, no body scanning, and no 360° photo shoots. Just smart algorithms plus data, data, data.
AI apparel application at its best?
On closer inspection, one could dispute that artificial intelligence is the exact opposite of algorithmic calculation. An algorithm describing a fixed sequence of defined steps that solve a problem. For example, a programmer implements a CAD program. Artificial intelligence generally refers to machine learning methods that are characterized by the fact that their results are not the result of a previously defined logic, but are "learned" from data. And it is precisely the availability of gigantic and increasingly specific, individual data sets – such as those for instance generated and leveraged by ZyseMe and FitAnalytics that enable the learning processes of machines, the creation of digital twins, and the interweaving of neural networks – to mention but a few technology topics building on AI.
Whatever the outcome of the discussion among scientific sectarians may be, here is the chance to at least reduce sustainable problems such as overproduction, understocking, and returns due to size sampling.
On February 21-23, 2022, the OECD (Organisation for Economic Cooperation and Development) will host the ‘International Conference on AI in Work, Innovation, Productivity, and Skills 2022’ and will feature leading voices from the technical, policy, business, academic, and civil society communities who will discuss insights on the adoption of AI in firms and the workplace, ethics of its use, and implications on skills, business dynamics, and productivity.
To attend this multi-stakeholder and evidence-based policy debate, users of the NEEDLE’S EYE can register below.