He builds IT systems that support future language research

“I like to say that I build tools for other researchers. They are the ones helping the world, and I help them help themselves,” says PhD student Niklas Deworetzki, whose research focuses on creating database systems for large datasets. 

“The goal is for researchers to be able to search for whatever they want, no matter how complex the question is, and get a result as quickly as possible. To do that the tools need to be efficient, but also understandable and usable.” 

Collaboration across disciplines 

Niklas Deworetzki’s research sits at the intersection of computer science and linguistics. As a computer scientist he therefore needs to collaborate closely with researchers in the humanities and social sciences. Together, they are currently working with a large collection of linguistic data containing more than 16 billion words in Språkbanken, a national infrastructure for language data. His task is to “upgrade” the search functionality of the database. This is done by building a new layer on top of the existing system, allowing it to understand, for example, more modern search languages. 

“There are already search tools for these collections of data, but a single query can take up to ten minutes. If a researcher realises the search conditions need to be adjusted, they have to wait just as long again. With our prototype, we can search the entirety of Swedish Wikipedia in around 100 milliseconds,” says Niklas Deworetzki. 

These searches enable researchers in the humanities and social sciences to study, for example, how political parties talk about different issues and how language changes over time. Recent developments in AI have had a major impact on linguistics, enabling automated language analysis or the identification of minority languages. But the computer science side has so far remained relatively unaffected, Niklas Deworetzki explains. 

“On the computer science side, we are essentially working with cold, hard math. You want a general system that always behaves correctly, where you can prove that it works the way it should. In the end, it often comes down to a small but well-thought-out piece of code, and that’s the kind of task AI struggles with.” 

Image

A newfound interest in language 

Niklas Deworetzki never imagined that he would work with languages, let alone language databases in an entirely new language. But moving from Germany to Sweden, and meeting colleagues from all over the world, has sparked a new fascination with languages. 

“I really enjoy learning about differences between languages and the funny expressions I hear from the other PhD students – like how Portuguese calls the calves ‘the potatoes of the leg’. After moving to Sweden, I’ve also gained a new appreciation for my own mother tongue, German. Just take the German word for sore muscles: ‘muscle hangover’. It’s amazing.” 

Niklas Deworetzki’s educational background has always been centered around computer science, ever since he taught himself programming in high school in order to create his own games in Minecraft. During his studies, his interest in research also grew. After working as a developer for a few years, he decided to return to academia. 

“I found industry work a bit repetitive and wanted to return to the creative environment at the university. In research, there are so many curious and passionate people,” he says. 

Strong support for PhD students 

Because the research field Niklas Deworetzki now works in was entirely new to him, the beginning of his PhD was a challenge. He needed to find his niche and understand which problems were truly interesting and important. Support from his supervisors and the research environment at the University of Gothenburg have therefore been crucial. 

“I quickly noticed that the University of Gothenburg genuinely cares about its PhD students. There are programmes, courses and support structures that give you room to grow. And the colleagues are incredibly valuable, I had no idea how many fantastic people I would meet when I moved here,” he says. 

What comes next is still open. He might stay in Sweden but could also imagine moving abroad. The one thing he knows for certain is that he wants to continue programming. 

“I want to keep working with programming languages and tools that help people."

Text: Natalija Sako