What an inspiring Research Demo Day we had!
The second edition of Research Demo Day was great! A short flashback:
The second edition of Research Demo Day was great! A short flashback:
Research Center: collaboration and results
Dennis Joosten (CTO) explained the audience about the Technology organisation of Info Support, why the Info Support Research Center was founded two years ago in the first place and what it has achieved so far.
“Innovation is one of the most important pillars at Info Support – and that is precisely why the company set up the Research Center two years ago. As part of our innovation program we invest in talented graduates from universities. We now have three lines of research that we focus on at our Research Center: Software Architecture, Data & AI and Software Development Methodologies. At Info Support, we are working on many innovative projects in these areas, and I can safely say that we are ahead of the market in this respect.
In Artificial Intelligence and Software Architecture, we have already worked alongside various universities to develop several programs. In these areas, we have entered partnerships with the University of Amsterdam and Vrije Universiteit Amsterdam, Utrecht University (UU), the University of Twente (UT), Radboud University (RU) in Nijmegen, and Wageningen University & Research (WUR).
I’m very pleased with the quality of the research carried out by the students who have completed their graduation project at the Info Support Research Center. All fifteen Master’s students who have done their research at Info Support’s Research Center scored an 8 or higher, and one even scored a 10. That 10 was for a student at the University of Amsterdam. It was only the third time that a 10 had been awarded there. That just goes to show that we are conducting high-quality research here at the Research Center.”
Update Research Area Artificial Intelligence
Joop Snijder (head of AI Research) explained the three main research topics of the Research Area Artificial Intelligence.“The first research topic is Dutch NLP, where NLP stands for Natural Language Processing. A great deal of research is being done in this area, but it almost always focuses on the English language. However, there are a number of specific problems related to the Dutch language. And our customers are mainly Dutch, which is why we need more research in this area.
Another main AI research topic is Explainable AI. ‘This is hugely important for us, because for the adoption of AI Machine Learning models it is vital that people can understand how a model actually learns and why it makes a certain decision. This means that we somehow have to open that black box. That’s the kind of thing that Explainable AI research focuses on.
The third research topic is Advanced Machine Learning, which includes the subject of computer vision.”
Thesis Talk: Automated privacy-preserving video processing and anonymized 3D scene reconstruction for surveillance video storage
Research student Lucía Conde Moreno, Master’s student Game and Media Technology at Utrecht University, has carried out her research in the field of Advanced Machine Learning under the supervision of Joop. She used a whole range of techniques based on previous research, with the aim of applying computer vision in a practical situation as a solution for one of the customers.
Among other things, she did this to find a solution to a question from Paaspop, a three-day music festival. During the festival, a lot of videos are made that the organization would like to save for an extended period, for future analysis. However, this raises a number of issues in terms of privacy legislation. The aim of the research was to develop an adequate, automated method that removes as many privacy-sensitive characteristics as possible from video recordings without rendering future analyses impossible research was to develop an adequate, automated method that removes as many privacy-sensitive characteristics as possible from video recordings without rendering future analyses impossible.
The anonymization of data and images is currently an important issue, especially in Europe due to the relatively strict GDPR legislation that has been introduced there. For example, one of the articles of the GDPR prohibits the long-term storage of this type of video, unless those videos have been stripped of all characteristics that can be traced back to people. Lucia Conde Moreno’s main research question was: what is the most effective method of removing all personal data from a video recording? She found that common methods developed for this purpose are not accurate or meticulous enough. For example, people often recognize blurred faces by a variety of other characteristics. Conde Moreno developed an ingenious new method for this problem, which was tested on Paaspop video material.
Update Research Area Software Architecture
Rinse van Hees (head of SA research) zoomed in on the main topics of the Research Area Software Architecture: the quality of software engineering and the resurgence of functional programming languages.
“A key research topic I’m working on is the quality of software engineering. When you develop software, you’re constantly building on other people’s work. So you need a guarantee that the quality of that work meets your requirements. In my view, that’s an area that hasn’t been getting the attention it deserves.
Our second key research topic is the resurgence of functional programming languages. Functional languages are among the oldest programming languages and have their roots in mathematics. Nowadays we have the computing power to program elegantly, so more and more functional programming constructions are appearing in the mainstream programming languages. This creates interesting interactions that are not always easy for the average developer to understand, because they have always been two very separate worlds.”
Thesis Talk: Code Quality Metrics for the Funtional Side of the Object-Oriented Language C#.
Bart Zuilhof, a Master’s student in Software Engineering at the University of Amsterdam, is studying at the Research Center under the supervision of Van Hees. He researched the rise of functional programming aspects in mainstream object-oriented programming languages, in this case C#. Since more and more functional programming elements are appearing in languages such as C#, new code metrics are needed to measure the quality of that software, he explained to Research Day visitors. Code metrics are specific standard measures for measuring the internal characteristics of a program code, such as the complexity of the code. The more complex the code, the more prone it is to errors. In his research, Zuilhof came up with a proposal for a number of new code metrics to measure the quality of the functional aspect of the current version of C#.
Nice to know: the paper of Bart has been accepted to Sattose 2019! This means he can present it at this Seminar on Advanced Techniques & Tools for Software Evolution in July in Bolzano.
Want to learn more about our research ambitions and topics?
Check out this site or contact us via email@example.com