From punch cards to artificial intelligence – from east to west

”When I studied at Kaunas University of Technology in Lithuania in the 1970s, there was only one computer in the entire university, and it took up the space of a large room. I had access to the computer once a week for two hours. That meant that I had to test the program in my head in advance. I learned to predict what would happen in each step of the program”, says Antanas Verikas, Professor of Image Processing with focus on artificial neural network technology at Halmstad University.

The idea that a computer can mimic the human nervous system, and thus learn to make the right decision, was launched in the 1950's. In the mid-1980's, Antanas Verikas, who at the time was studying character and pattern recognition, was fascinated by this area called artificial neural network (neuron = nerve cell). A neural network is a structure that is trained to solve a particular problem. For example, it can recognise patterns in a large amount of data or it can create predictions. This can for instance be used to estimate whether a person is healthy or ill. Input is a set of health parameters, such as pulse and blood pressure. Output can in this case be an answer to if the person is healthy – a yes or a no. However, neural networks must be trained to provide the correct output, i.e. make the right decision.

”Today, the artificial neural network models can be very complex and handle thousands of input variables. In my research, I have developed methods for assessing and selecting these variables”, says Antanas Verikas.

A specific example of when these models can be used is in predicting muscle fatigue during exercise. Antanas Verikas and his colleagues at the Center for Applied Intelligent Systems Research (CAISR) measured in a research project cyclists' EMG signals, blood lactate concentration levels and oxygen uptake. The data ​​was used to create models to evaluate the cyclists' physical ability under increased strain.

”This can, for instance, be used by an elite cyclist when setting up his or her training. The models provide important information about when a potential muscle fatigue occurs”, says Antanas Verikas.

Research on behalf of the Soviet military

Antanas Verikas interest in hiking and climbing was stirred in young years. Every summer was spent in the Pamir, Tian Shan or Caucasus mountain ranges.

He grew up in Lithuania when the country was occupied by the Soviet Union. Soviet controlled large parts of the society – as well as the university's research.

”Most research projects at Kaunas Technical University were at that time commissioned by the Soviet military. During my time as a PhD student, we developed the navy's database for submarines. They wanted to transfer text on paper about the submarines to their database. Our mission was to create hardware and software for this, equivalent to today's scanner.”

Antanas Verikas developed the software algorithms. Unlike today, all programming was done by hand, on paper. Paper strips were used to transfer programs to a computer, with holes representing the program code of ones and zeros. The strips, that could be over 30 meters long, were replaced by punch cards with the equivalent hole coding in the late 1970's.

”We had to write very effective programs because the memory of the computer was so limited at that time. The programs could not exceed 25 kb, if I remember correctly.”

The application of the research results was classified. As a PhD student, Antanas Verikas went to Leningrad twice a year to report to the military what the research group had achieved.

”I was interviewed by a group of senior military personnel. I had to keep my head cool to answer, defend and explain our research, says Antanas Verikas and chuckles at the memory.”

As an active researcher, you could not talk to fellow scholars outside of the Soviet Union. And you were not permitted to travel westward.

”I tried to go to Italy as a guest researcher in 1985, but the leaders in Moscow did not allow it. In 1991, the Soviet Union dissolved and the borders were opened. It felt amazing!”

From paper and print quality to health innovation

Through one of his doctoral students at Kaunas University of Technology, Antanas Verikas established a contact with researchers at Halmstad University. And through a scholarship from the Swedish Institute, Antanas Verikas came to Halmstad in October 1992.

”I planned to stay for nine months, but 26 years have passed”, he says with a smile.

Antanas Verikas and his research team started a collaboration with several of Sweden's paper mills. They were interested in knowing which variables resulted in the best paper and print quality.

”As a spin-off, we also worked with banknotes. We developed hardware and software to assess the quality of the details on the banknote. The purpose was to determine if the printed notes were of sufficient quality to prevent copies. We delivered ready-made systems for this to, for example, the Bank of England in the early 2000's.”

When CAISR was initiated in 2010, two application areas were defined: health technology and intelligent vehicles.

”A year later, we received funding from the KK Foundation. Since then our research center has grown every year”, says Antanas Verikas.

2018 is Antanas Verikas final year as a researcher. When he retires by the end of the year, he wants to focus on his hobbies: hiking and mountain climbing. A certain longing for home can also be discerned:

”I want to travel! To visit Burma, Cambodia and Vietnam. Madeira is also fascinating. And I want to spend more time in Lithuania.”

Text: Louise Wandel
Photo: Joachim Brink