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.
I am also working with artificial intelligence
The academic chemometrics veteran Rasmus Bro is not at all negative about the hype surrounding AI and machine learning. Data analysis is growing. Analytics is playing an increasingly important role in business and society. It’s a good thing. He wouldn’t hesitate to state that he himself is working with ‘artificial intelligence’.
Even if you have dedicated your life to jazz, you shouldn’t be sad, angry or arrogant, when rock or pop dominates and fill up the concert halls. It’s all music.
This metaphor illustrates the view point of Rasmus Bro, harmonica player in a part time professional blues band and professor at Copenhagen University at the section for Chemometrics and Analytical Technology. He points out the resemblance and the common denominator between all members of the analytics community.
“We are the old established guys now, and we should applaud and embrace all the new focus on analytics,” says Rasmus Bro. “When chemometrics was young, it was the other way around. When mathematicians listened to pioneers in chemometrics like Harald Martens some decades ago, they saw a new exciting way to approach the same field of models and analysis. The difference was more between boring and exciting than in basic understanding. Now we should stay positive, even though machine learning and deep learning is running with the glory, the fame and the funding right now. I am ready to do what I do and call it machine learning, if it strengthens our funding and reputation.”
Rasmus Bro remembers how lean, Six Sigma and Business Process Re-engineering made headlines decades ago. A lot of money were waisted on projects in these hyped management disciplines and in the field of genomics. The same could be said for the current focus on AI. But history moves on. 25 years ago everybody talked about neural networks but little came out of that at the time.
Crucial time in history
“Words like deep learning and machine learning will probably also disappear again,” he says. “The focus on data analysis is the important aspect, and it is growing right now. We are probably at a crucial time in history in terms of these technologies, and right now we see a boom in concepts and competing approaches. Big Data, IoT and machine learning holds the potential to revolutionize, but I think it will settle down and become more like infrastructure. It will become more standardized. To me it makes sense to compare with cars. At the earliest days you couldn’t predict where it would go. Later all the standards in fuel, engines and roads left the car basically unchanged in 70 years.”
Rasmus Bro does not consider himself to be an especially gifted futurist. He didn’t foresee the Orwellian degree of surveillance, which he feels is the reality right now in modern societies. He is a highly specialized scientist in chemometrics and as such in the field of data analysis. This discipline is more than 100 years old, and it will continue to evolve. At least Rasmus Bro continue to do his research projects in the tight relationships between the university and the sponsoring companies. His specialty is multi-way analysis, which he describes as “getting information from a box filled with data”. Fueled by new capabilities in measurement of chemical characteristics chemometrics is all about speeding up the process from measurement to knowledge.
Speeding up chemometrics
“Chemometrics as a discipline in science is connected to the ability to measure and the ability to process data. Earlier you had measurement in a lab and days, weeks or months later you had the outcome. Today you measure and get sophisticated results right away. This is valuable in pharma, chemistry, food and other industries looking for continuous manufacturing and integrated quality control. There is still a lot to gain from the integration of analytics directly to the process,” he says.
According to Rasmus Bro the first wave of chemometrics originated in Sweden with Svante Wold as the key person and later Norway with Harald Martens played a significant role. Rasmus Bro sees mathematics playing a key role going forward in terms of multi-way analysis and modelling of complex data. One of the characteristics with chemometrics is the emphasis on solving real problem through science. At University of Copenhagen Rasmus Bro and his peers work in projects funded by corporations in food, pharma and chemistry.
“We are pleased to work closely with real life problems. It may not give us fame and glory in academia, but we are driven by the real challenges in our field. It may be the candy company having problems with the consistency in marzipan or the pharmaceutical company going for increased quality control in every tablet leaving the factory. To the company it is valuable problem solving, and we are often able to combine the solution in real life with basic scientific research,” he says.
Beer and food
One of the projects on the drawing board in Copenhagen is about beer and food. Carlsberg has had a historical role in supporting Copenhagen University in working with food, and now the chemometricians see a much deeper analysis of flavors and traits in beer to match up with different food.
“We can describe beer in much more detail than is suggested by overall categories like ‘IPA’ or ‘lager’. This knowledge could be utilized in fine dining to make a match of all the diversity in beers with all the diversity in food. This knowledge could further be matched with text analysis to find out how human beings describe the beer experience. To me it is quite obvious that the capabilities in chemometrics have a lot of potential in a whole range of sexy applications,” he says.
Sharing creates knowledge
“My job is not to capitalize on this science. Our task is to produce more knowledge and educate young people. On that mission it has proven very valuable to share our results and findings. We started out in this fashion in 1994, and it has become our way to work. We share everything we can share, and in turn we are building a great network with other scientific hotspots and with companies. At first it was not a strategy. We did it because we were proud of our accomplishments. Now we can see how valuable it has been to connect and cooperate globally,” says Rasmus Bro.
One of the very crucial traits in chemometrics is of course the connection to chemistry. Combining domain knowledge and advanced analytics is the name of the game, but Rasmus Bro sees the very same trend in AI.
“On the one hand you see a too big fascination with AI, and I think a lot of money will be spend with no result in that arena. On the other hand the AI space is run in consortium with companies like Google sharing results and models generously. If you look at the best in deep learning and machine learning, you will see willingness and eagerness to exploit and build on domain knowledge. It is not far from how we work and how chemometrics makes results,” says Rasmus Bro.
He emphasizes that chemometrics is a fairly small field in science, and everybody basically knows everybody. Also companies like Camo Analytics and others are deeply rooted in this environment.
“Lean, Six Sigma and Business Process Re-engineering made headlines decades ago. A lot of money were waisted on projects in these hyped management disciplines and in the field of genomics. The same could be said for the current focus on AI. But history moves on. 25 years ago everybody talked about neural networks but little came out of that at the time.”
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.