Our AI Journey

Advancing Scientific Summaries through Transformer Language Models with Juraj

Juraj explores the synergy between AI and human perspectives, the transformative potential of his project and shares details about his collaborations.


Juraj Vladika

PhD Student | Software Campus Participant

TU München

Hello Juraj, tell us about yourself!

Hi! I am Juraj, a second-year PhD student of Computer Science at the Technical University of Munich. My field of study is Natural Language Processing. I come from Croatia, where I completed my bachelor’s and master’s degrees at the University of Zagreb. Having studied in four countries, I enjoy international environments, exploring new places, and meeting new people. In my free time, I like being in nature, reading, language learning, and live music.

What fascinates you about your field of study?

I chose to study computer science and later specialized in artificial intelligence because I have always found it amazing how we can use machines to help us solve challenging tasks. I have also been fascinated since childhood by languages and with how people communicate. I believe language is the most powerful tool people have, and analyzing it can give us insight into the human mind. The field of Natural Language Processing, which lies at the intersection of these two worlds, was a perfect fit for me.

What is your project about and with whom are you collaborating on it?

My project is on scientific claim verification and is in collaboration with Digital Science. In the project, the goal is to investigate and develop systems that, for a given claim related to science, aim to find relevant evidence for it in scientific articles and verify its validity. This involves working with information retrieval over large knowledge bases and knowledge graphs, discovering arguments in text, and using the reasoning capabilities of language models to produce predictions.

How could your project be applied in the future?

There are multiple applications of this project and its findings. The first is helping scientists conduct their research work by automating the process of finding valuable resources and evidence for their research hypotheses. Additionally, the system could be used to curb the spread of scientific misinformation by checking the truthfulness of trending online claims. Finally, an important application is using claim verification to improve the factuality of large language models by evaluating their responses and reducing the so-called hallucinations in their output.

How is the Software Campus program contributing to your personal development?

The program has offered us participation in interactive workshops on various topics such as leadership, teamwork, change management, and innovative thinking. These workshops helped my personal development by focusing on the soft skills that are often neglected in the tech world but are of utter importance in daily life. Additionally, we have had the opportunity to network with people of similar interests all across Germany and exchange experience on our projects and work. This fosters stronger collaboration and integration of German researchers working in computer science.

What collaboration and network opportunities have you had during the program?

Along with the mentioned workshops, we have also had the opportunity to attend events such as the Software Campus Summit and Holtzbrinck Innovation Days. These events allowed us to present our project to a broader audience and hear about the latest AI developments in both academia and industry. Active collaboration between academia and industry is essential to make progress by ensuring that research work in academia will have an applied setting in the future while encouraging the industry to employ the latest advancements from AI research in their products and services.

Back to overview