Industrial analytics thought leadership-series.
How will industrial analytics, Industry 4.0, IIoT and digital transformation change businesses? And ultimately the world? We ask industry leaders to share their view on these technologies, how we use them, and where they will take us.
“It is extremely difficult to automate in full. There are variations in materials, components and processes. The variation is huge, and I cannot imagine totally automated manufacturing except for very simple products. I don’t believe much in the concept of continuous manufacturing of pharmaceutical products like proteins and other polymers. Here you will need more advanced models,” says Svante Wold.
The retired chemometrics veteran with an impressive lifelong achievement within industrial analytics emphasizes the progress in software. Analytics has expanded in breadth and depth, and the methodology behind chemometrics has proven its worth to problems in other industries. The automation and improvements have been going on for years and the software has penetrated industry barriers and has proven applicable to any data.
Can’t stand alone
“Like AI these tools have proven themselves on any dataset. Chemometrics is also AI. But from my viewpoint neither can stand alone.”
“It’s a big mistake to think that the truth lies in data. It is only a part of the truth and a part of the future. Data as the only solution is stupid. This all data centric approach has the same limitation, and it goes for both AI and chemometrics,” says Svante Wold.
The skepticism towards AI and continuous manufacturing has to do with all the unknowns still out there. The overly enthusiastic prophets of AI reduce complexity in real world problems.
“Data is very good at many things. AI will not cure cancer, but it may very well help us diagnose,” says Svante Wold. “I wouldn’t say, that I am furious about all the talk of AI. But I am irritated. I think more business people should learn some simple data analysis to understand what it is all about.”
The Godfather of chemometrics
Svante Wold was a young PhD at Umeå University, Sweden, in 1971, when he invented the word ‘chemometrics’ for a grant application. One of the effects of having new electronic devices in the ’70s was the production of more data. So with the new instruments also came the need for analytical tools to crunch the tide of emerging data. One of the main characteristics was the presence of many more variables than observations.
“Chemometrics is the art of extracting chemically relevant information from data produced in chemical experiments,” wrote Svante Wold. The same ideas of analyzing multidimensional chemical data, through the application of statistics and applied math, via computers, was shared by other chemists in the west coast of the USA. Svante Wold travelled frequently there and in the ’70s The International Chemometrics Society was founded by Svante and Bruce Kowalski and their scientific groups.
PLS Regression in the ’70s and ’80s
The early interest of chemometrics were mainly about pattern recognition and classification. In the ’70s Svante started to work with the PLS model, and the meeting with Norwegian chemist Harald Martens in Oslo became very fruitful for the development of PLS Regression. The method proved very successful and promising.
“We had a wonderful time,” remembers Svante Wold. “Those years between 1980 and 1984 were among the happiest—scientifically—in our professional lives.”
The PLS regression framework was very practical, useful and valuable. In an industry where there was a need for analytical tools capable of dealing with multicollinearity, missing values, and large data sets, PLS Regression fulfilled all those needs. Svante Wold became a premier evangelist for the multivariate analysis within industrial analytics.
Svante, Nouna and the female Tesla
“I met Svante at a conference in New Hampshire. He talked about PLS. We were married one year later,” says Nouna Kettaneh, who has been working alongside Svante Wold in designing and developing software for advanced industrial analytics. They share the skeptical stance on AI:
“AI has come a long way, but it has to be linked to humans and it has to be complemented,” says Nouna Kettaneh. “The human brain will stay on main stage, and AI and analytics will be a powerful tool. We need to automate, but we must acknowledge the need to understand and correct the processes.”
Nouna and Svante drive a Tesla, which they describe as a computer on wheels.
“I am really happy for her (the car is female according to Nouna). I am a lousy driver, and the car helps me quite a lot,” says Nouna Kettaneh.
“It’s a nice car, but it is not complete. I do not see cars become selfdriving,” says Svante Wold.
The parallel is evident, when talking to Svante and Nouna. Industrial analytics should aspire to be an ever more advanced tool box for the domain scientists in development and manufacturing. The car is a tool that should support the human being as the driver, not taking over.
Find the right solution for your analytics needs.
Get in touch if you have questions about our products, platform, how to get started or how best to address your analytics needs.