A mathematician creates privacy

In his doctoral research, Thanh-Long Tran is looking for ways to ensure that anonymous data contain no paths back to individuals or their details. His research project in machine learning is part of collaboration between the University of Helsinki and Nokia Bell Labs.

“Researchers’ work is about brainstorming based on theoretical thinking. First comes an idea, and when it works on paper, you test it. Validated results create new research that benefits machine learning,” says Doctoral Researcher Thanh-Long Tran

Tran works as a researcher and is writing his doctoral thesis in a collaborative project carried out by the University of Helsinki and Nokia Bell Labs, Nokia’s research and development arm. His topic, differential privacy, is a new research field in machine learning. It focuses on protecting privacy in the processing of large datasets, which can concern, for example, people’s health. 

In practice, Tran investigates how techniques such as randomness and what is known as noise must be incorporated into computing systems to ensure that anonymised data cannot be later connected to individuals. In other words, the goal is to ensure that computers cannot retrace the same path of zeros and ones to indicate which details relate to specific individuals.  

Tran finds it important for research to have significance for both himself and society, as well as the global field of machine learning. What fascinates him in research is the creation of new knowledge. 

“New knowledge can be either incremental or revolutionary. Incremental refines prior research, while revolutionary discovers an entirely new method, for instance. That’s the ultimate dream, but it’s not easy to achieve.” 

From physics to mathematics

Thanh-Long Tran, born in Vietnam, was supposed to become a physicist. In 2020 the European Physics Olympiad (EuPhO) welcomed participants from around the world, and Tran took part while studying in general upper secondary school. He won a silver medal, which opened the door to studying at the University of Helsinki on a scholarship. 

In Helsinki, Tran noticed that mathematical problem-solving was more fascinating than theoretical physics. 

“My brain is fond of straightforward and transparent logic and probability theory.” 

Tran completed his bachelor’s and master’s degrees in four years. In the summer of 2025, he started as a doctoral researcher under the supervision of Professor Antti Honkela and Senior Machine Learning Researcher, Docent Antti Koskela of Nokia Bell Labs. 

Originally, Tran planned to find a job in the financial sector immediately after graduating with a master’s degree. However, Professor Honkela suggested embarking on doctoral research. 

“I’ve noticed that research suits me. I have not yet decided what I will do after completing my doctoral education. I believe a doctoral degree will help me either find employment in the private sector or advance my academic career.” 

I believe a doctoral degree will help me either find employment in the private sector or advance my academic career.
Fresh ideas from different environments

The collaboration between the University of Helsinki and Nokia Bell Labs began in 2017, when the latter selected the University for its Distinguished Academic Partnership programme. For 90 years, Nokia Bell Labs has laid the groundwork for information and telecommunication networks and digital devices. Their research has garnered eight Nobel Prizes.  

Tran believes that research collaboration is a concrete way of tackling global phenomena, in this case the reliability and safety of machine learning. It ensures the quality of research, its applicability to practice and, ultimately, user-friendliness. In practice, collaboration has meant bringing various ideas and mindsets together. 

“Nokia Bell Labs researchers work in a different environment than we do at the University. When I end up at an impasse in my research, they often have another perspective on the matter. Together, we then solve the problem.” 

A plurality of perspectives among researchers speeds up research as a process, and when breakthroughs occur, they directly enhance the ways in which data are processed and utilised in society. 

“We only need one idea that works. If that happens, it’s a success.” 

Doctoral graduates in professional life and businesses
  • Doctoral education provides professional skills alongside competencies for the labour market. More than two-thirds of doctoral graduates work outside universities, and this share is growing steadily. 
  • The number of doctoral graduates is growing particularly in the private sector and among entrepreneurs: in 2022 the share of doctoral graduates working in this group was 35.5%, up from 28.8% in 2015. 
  • In the ongoing doctoral education pilot in 2024–2027, Finnish universities will train 1,000 new doctoral graduates and reform the practices of doctoral education. A total of more than 10,000 applications from graduates who will intensively collaborate with businesses were received for the pilot. 
  • The unemployment rate of doctoral graduates is lower than that for other levels of education, and their salaries are higher. The difference compared to master's degree holders is 16% and 48% compared to graduates from universities of applied sciences. 
  • Of the scholarly articles produced at Finnish higher education institutions, 97% originate in universities. The same applies to patents and invention disclosures. 

Source: Council of Rectors of Finnish Universities (UNIFI)