Chiara Di Bonaventura on her experience at the International Semantic Web Research Summer School

30th August 2023 | News

News > Chiara Di Bonaventura on her experience at the International Semantic Web Research Summer School

Chiara Di Bonaventura, PhD student (2022 cohort) from the UKRI Centre for Doctoral Training in Safe and Trusted AI (the STAI CDT), recently attended the International Semantic Web Research Summer School (ISWS) which was held in Bertinoro, Italy in July 2023. The ISWS is an intensive one-week summer school with a focus on collaboration, teamwork, creativity and networking. 

The ISWS is a unique experience. As Chiara explains, “It is very different from any other summer school, that’s why I really enjoyed it! It combined the usual lectures you can typically find in summer schools with research group projects and presentations to help young researchers  to develop both their knowledge but also important skills such as communication and working as part of a team. I personally think this ‘learning by doing’ approach was the main highlight of the summer school.” 

Alongside lectures and research group projects, the ISWS also focuses on fun and socialising, with various activities to help foster a friendly environment. Chiara describes a typical day at the ISWS here, “We used to start the day with a yoga session all together, followed by a brainstorming session for the group project. Then, we usually had keynote speakers for the rest of the morning and team work on the group project in the afternoon. Throughout the day, there were opportunities for networking and social activities.” 

Morning yoga at the ISWS

The ISWS also offers early-career researchers an opportunity to present their work and learn from others through poster presentations, and the group project provides a starting point for future collaboration with peers.  

The project Chiara worked on investigated whether Knowledge Graphs (KGs) add any value to the performance of Large Language Models. There have been discussions that they help the performance of LLMs for reducing hallucination, providing explanations, and zero-shot learning. Chiara’s group focused on the context of Creative AI and chose the task of story completion (by asking the model to choose among different alternatives to complete a natural language story). Their results suggested that enhancing LLMs with KGs actually hinders performance on this task, raising interesting avenues for future research to explore the reasons for this.  

Chiara gained a lot from attending the summer school and said, “What I take home from this experience is the opportunity to **actively** learn from outstanding researchers in the field of Semantic Web not only through lectures but especially through discussions, poster presentations and teamwork. I am still in contact with my research group, and we would like to extend the research project we started at ISWS 2023”.